fix

* Confluence fixes

* Small tweak

* Address greptile comments

.github/CODEOWNERS

@@ -0,0 +1 @@
+* @onyx-dot-app/onyx-core-team

.github/workflows/nightly-scan-licenses.yml

@@ -53,24 +53,90 @@ jobs:
           exclude: '(?i)^(pylint|aio[-_]*).*'
           
       - name: Print report
-        if: ${{ always() }}
+        if: always()
         run: echo "${{ steps.license_check_report.outputs.report }}"
       
       - name: Install npm dependencies
         working-directory: ./web
         run: npm ci
-        
-      - name: Run Trivy vulnerability scanner in repo mode
-        uses: aquasecurity/trivy-action@0.28.0
-        with:
-          scan-type: fs
-          scanners: license
-          format: table
-#           format: sarif
-#           output: trivy-results.sarif
-          severity: HIGH,CRITICAL
 
-#       - name: Upload Trivy scan results to GitHub Security tab
-#         uses: github/codeql-action/upload-sarif@v3
+        # be careful enabling the sarif and upload as it may spam the security tab
+        # with a huge amount of items. Work out the issues before enabling upload.       
+#       - name: Run Trivy vulnerability scanner in repo mode
+#         if: always()
+#         uses: aquasecurity/trivy-action@0.29.0
 #         with:
-#           sarif_file: trivy-results.sarif
+#           scan-type: fs
+#           scan-ref: .
+#           scanners: license
+#           format: table
+#           severity: HIGH,CRITICAL
+# #           format: sarif
+# #           output: trivy-results.sarif
+# 
+# #       - name: Upload Trivy scan results to GitHub Security tab
+# #         uses: github/codeql-action/upload-sarif@v3
+# #         with:
+# #           sarif_file: trivy-results.sarif
+
+  scan-trivy:
+    # See https://runs-on.com/runners/linux/
+    runs-on: [runs-on,runner=2cpu-linux-x64,"run-id=${{ github.run_id }}"]
+      
+    steps:
+    - name: Set up Docker Buildx
+      uses: docker/setup-buildx-action@v3
+
+    - name: Login to Docker Hub
+      uses: docker/login-action@v3
+      with:
+        username: ${{ secrets.DOCKER_USERNAME }}
+        password: ${{ secrets.DOCKER_TOKEN }}
+
+    # Backend
+    - name: Pull backend docker image
+      run: docker pull onyxdotapp/onyx-backend:latest
+
+    - name: Run Trivy vulnerability scanner on backend
+      uses: aquasecurity/trivy-action@0.29.0
+      env:
+        TRIVY_DB_REPOSITORY: 'public.ecr.aws/aquasecurity/trivy-db:2'
+        TRIVY_JAVA_DB_REPOSITORY: 'public.ecr.aws/aquasecurity/trivy-java-db:1'
+      with:
+        image-ref: onyxdotapp/onyx-backend:latest
+        scanners: license
+        severity: HIGH,CRITICAL
+        vuln-type: library
+        exit-code: 0  # Set to 1 if we want a failed scan to fail the workflow
+
+    # Web server
+    - name: Pull web server docker image
+      run: docker pull onyxdotapp/onyx-web-server:latest
+          
+    - name: Run Trivy vulnerability scanner on web server
+      uses: aquasecurity/trivy-action@0.29.0
+      env:
+        TRIVY_DB_REPOSITORY: 'public.ecr.aws/aquasecurity/trivy-db:2'
+        TRIVY_JAVA_DB_REPOSITORY: 'public.ecr.aws/aquasecurity/trivy-java-db:1'
+      with:
+        image-ref: onyxdotapp/onyx-web-server:latest
+        scanners: license
+        severity: HIGH,CRITICAL
+        vuln-type: library
+        exit-code: 0
+
+    # Model server
+    - name: Pull model server docker image
+      run: docker pull onyxdotapp/onyx-model-server:latest
+
+    - name: Run Trivy vulnerability scanner
+      uses: aquasecurity/trivy-action@0.29.0
+      env:
+        TRIVY_DB_REPOSITORY: 'public.ecr.aws/aquasecurity/trivy-db:2'
+        TRIVY_JAVA_DB_REPOSITORY: 'public.ecr.aws/aquasecurity/trivy-java-db:1'
+      with:
+        image-ref: onyxdotapp/onyx-model-server:latest
+        scanners: license
+        severity: HIGH,CRITICAL
+        vuln-type: library
+        exit-code: 0

.github/workflows/pr-integration-tests.yml

@@ -99,7 +99,7 @@ jobs:
           DISABLE_TELEMETRY=true \
           IMAGE_TAG=test \
           DEV_MODE=true \
-          docker compose -f docker-compose.multitenant-dev.yml -p danswer-stack up -d
+          docker compose -f docker-compose.multitenant-dev.yml -p onyx-stack up -d
         id: start_docker_multi_tenant
 
       # In practice, `cloud` Auth type would require OAUTH credentials to be set.
@@ -108,12 +108,13 @@ jobs:
           echo "Waiting for 3 minutes to ensure API server is ready..."
           sleep 180
           echo "Running integration tests..."
-          docker run --rm --network danswer-stack_default \
+          docker run --rm --network onyx-stack_default \
             --name test-runner \
             -e POSTGRES_HOST=relational_db \
             -e POSTGRES_USER=postgres \
             -e POSTGRES_PASSWORD=password \
             -e POSTGRES_DB=postgres \
+            -e POSTGRES_USE_NULL_POOL=true \
             -e VESPA_HOST=index \
             -e REDIS_HOST=cache \
             -e API_SERVER_HOST=api_server \
@@ -143,24 +144,28 @@ jobs:
       - name: Stop multi-tenant Docker containers
         run: |
           cd deployment/docker_compose
-          docker compose -f docker-compose.multitenant-dev.yml -p danswer-stack down -v
+          docker compose -f docker-compose.multitenant-dev.yml -p onyx-stack down -v
 
+      # NOTE: Use pre-ping/null pool to reduce flakiness due to dropped connections
       - name: Start Docker containers
         run: |
           cd deployment/docker_compose
           ENABLE_PAID_ENTERPRISE_EDITION_FEATURES=true \
           AUTH_TYPE=basic \
+          POSTGRES_POOL_PRE_PING=true \
+          POSTGRES_USE_NULL_POOL=true \
           REQUIRE_EMAIL_VERIFICATION=false \
           DISABLE_TELEMETRY=true \
           IMAGE_TAG=test \
-          docker compose -f docker-compose.dev.yml -p danswer-stack up -d
+          INTEGRATION_TESTS_MODE=true \
+          docker compose -f docker-compose.dev.yml -p onyx-stack up -d
         id: start_docker
 
       - name: Wait for service to be ready
         run: |
           echo "Starting wait-for-service script..."
 
-          docker logs -f danswer-stack-api_server-1 &
+          docker logs -f onyx-stack-api_server-1 &
 
           start_time=$(date +%s)
           timeout=300  # 5 minutes in seconds
@@ -190,15 +195,24 @@ jobs:
           done
           echo "Finished waiting for service."
 
+      - name: Start Mock Services
+        run: |
+          cd backend/tests/integration/mock_services
+          docker compose -f docker-compose.mock-it-services.yml \
+            -p mock-it-services-stack up -d
+
+      # NOTE: Use pre-ping/null to reduce flakiness due to dropped connections
       - name: Run Standard Integration Tests
         run: |
           echo "Running integration tests..."
-          docker run --rm --network danswer-stack_default \
+          docker run --rm --network onyx-stack_default \
             --name test-runner \
             -e POSTGRES_HOST=relational_db \
             -e POSTGRES_USER=postgres \
             -e POSTGRES_PASSWORD=password \
             -e POSTGRES_DB=postgres \
+            -e POSTGRES_POOL_PRE_PING=true \
+            -e POSTGRES_USE_NULL_POOL=true \
             -e VESPA_HOST=index \
             -e REDIS_HOST=cache \
             -e API_SERVER_HOST=api_server \
@@ -208,6 +222,8 @@ jobs:
             -e CONFLUENCE_USER_NAME=${CONFLUENCE_USER_NAME} \
             -e CONFLUENCE_ACCESS_TOKEN=${CONFLUENCE_ACCESS_TOKEN} \
             -e TEST_WEB_HOSTNAME=test-runner \
+            -e MOCK_CONNECTOR_SERVER_HOST=mock_connector_server \
+            -e MOCK_CONNECTOR_SERVER_PORT=8001 \
             onyxdotapp/onyx-integration:test \
             /app/tests/integration/tests \
             /app/tests/integration/connector_job_tests
@@ -229,13 +245,13 @@ jobs:
         if: always()
         run: |
           cd deployment/docker_compose
-          docker compose -f docker-compose.dev.yml -p danswer-stack logs --no-color api_server > $GITHUB_WORKSPACE/api_server.log || true
+          docker compose -f docker-compose.dev.yml -p onyx-stack logs --no-color api_server > $GITHUB_WORKSPACE/api_server.log || true
 
       - name: Dump all-container logs (optional)
         if: always()
         run: |
           cd deployment/docker_compose
-          docker compose -f docker-compose.dev.yml -p danswer-stack logs --no-color > $GITHUB_WORKSPACE/docker-compose.log || true
+          docker compose -f docker-compose.dev.yml -p onyx-stack logs --no-color > $GITHUB_WORKSPACE/docker-compose.log || true
 
       - name: Upload logs
         if: always()
@@ -249,4 +265,4 @@ jobs:
         if: always()
         run: |
           cd deployment/docker_compose
-          docker compose -f docker-compose.dev.yml -p danswer-stack down -v
+          docker compose -f docker-compose.dev.yml -p onyx-stack down -v

.github/workflows/pr-playwright-tests.yml

@@ -1,6 +1,6 @@
-name: Run Chromatic Tests
+name: Run Playwright Tests
 concurrency:
-  group: Run-Chromatic-Tests-${{ github.workflow }}-${{ github.head_ref || github.event.workflow_run.head_branch || github.run_id }}
+  group: Run-Playwright-Tests-${{ github.workflow }}-${{ github.head_ref || github.event.workflow_run.head_branch || github.run_id }}
   cancel-in-progress: true
 
 on: push
@@ -198,43 +198,47 @@ jobs:
           cd deployment/docker_compose
           docker compose -f docker-compose.dev.yml -p danswer-stack down -v
 
-  chromatic-tests:
-    name: Chromatic Tests
+# NOTE: Chromatic UI diff testing is currently disabled.
+# We are using Playwright for local and CI testing without visual regression checks.
+# Chromatic may be reintroduced in the future for UI diff testing if needed.
 
-    needs: playwright-tests
-    runs-on:
-      [
-        runs-on,
-        runner=32cpu-linux-x64,
-        disk=large,
-        "run-id=${{ github.run_id }}",
-      ]
-    steps:
-      - name: Checkout code
-        uses: actions/checkout@v4
-        with:
-          fetch-depth: 0
+# chromatic-tests:
+#   name: Chromatic Tests
 
-      - name: Setup node
-        uses: actions/setup-node@v4
-        with:
-          node-version: 22
+#   needs: playwright-tests
+#   runs-on:
+#     [
+#       runs-on,
+#       runner=32cpu-linux-x64,
+#       disk=large,
+#       "run-id=${{ github.run_id }}",
+#     ]
+#   steps:
+#     - name: Checkout code
+#       uses: actions/checkout@v4
+#       with:
+#         fetch-depth: 0
 
-      - name: Install node dependencies
-        working-directory: ./web
-        run: npm ci
+#     - name: Setup node
+#       uses: actions/setup-node@v4
+#       with:
+#         node-version: 22
 
-      - name: Download Playwright test results
-        uses: actions/download-artifact@v4
-        with:
-          name: test-results
-          path: ./web/test-results
+#     - name: Install node dependencies
+#       working-directory: ./web
+#       run: npm ci
 
-      - name: Run Chromatic
-        uses: chromaui/action@latest
-        with:
-          playwright: true
-          projectToken: ${{ secrets.CHROMATIC_PROJECT_TOKEN }}
-          workingDir: ./web
-        env:
-          CHROMATIC_ARCHIVE_LOCATION: ./test-results
+#     - name: Download Playwright test results
+#       uses: actions/download-artifact@v4
+#       with:
+#         name: test-results
+#         path: ./web/test-results
+
+#     - name: Run Chromatic
+#       uses: chromaui/action@latest
+#       with:
+#         playwright: true
+#         projectToken: ${{ secrets.CHROMATIC_PROJECT_TOKEN }}
+#         workingDir: ./web
+#       env:
+#         CHROMATIC_ARCHIVE_LOCATION: ./test-results

.github/workflows/pr-python-connector-tests.yml

@@ -74,7 +74,9 @@ jobs:
           python -m pip install --upgrade pip
           pip install --retries 5 --timeout 30 -r backend/requirements/default.txt
           pip install --retries 5 --timeout 30 -r backend/requirements/dev.txt
-
+          playwright install chromium
+          playwright install-deps chromium
+          
       - name: Run Tests
         shell: script -q -e -c "bash --noprofile --norc -eo pipefail {0}"
         run: py.test -o junit_family=xunit2 -xv --ff backend/tests/daily/connectors

.github/workflows/pr-python-model-tests.yml

@@ -1,18 +1,29 @@
-name: Connector Tests
+name: Model Server Tests
 
 on:
   schedule:
     # This cron expression runs the job daily at 16:00 UTC (9am PT)
     - cron: "0 16 * * *"
-
+  workflow_dispatch:
+    inputs:
+      branch:
+        description: 'Branch to run the workflow on'
+        required: false
+        default: 'main'
+        
 env:
   # Bedrock
   AWS_ACCESS_KEY_ID: ${{ secrets.AWS_ACCESS_KEY_ID }}
   AWS_SECRET_ACCESS_KEY: ${{ secrets.AWS_SECRET_ACCESS_KEY }}
   AWS_REGION_NAME: ${{ secrets.AWS_REGION_NAME }}
 
-  # OpenAI
+  # API keys for testing
+  COHERE_API_KEY: ${{ secrets.COHERE_API_KEY }}
+  LITELLM_API_KEY: ${{ secrets.LITELLM_API_KEY }}
+  LITELLM_API_URL: ${{ secrets.LITELLM_API_URL }}
   OPENAI_API_KEY: ${{ secrets.OPENAI_API_KEY }}
+  AZURE_API_KEY: ${{ secrets.AZURE_API_KEY }}
+  AZURE_API_URL: ${{ secrets.AZURE_API_URL }}
 
 jobs:
   model-check:
@@ -26,6 +37,23 @@ jobs:
       - name: Checkout code
         uses: actions/checkout@v4
 
+      - name: Login to Docker Hub
+        uses: docker/login-action@v3
+        with:
+          username: ${{ secrets.DOCKER_USERNAME }}
+          password: ${{ secrets.DOCKER_TOKEN }}
+
+      # tag every docker image with "test" so that we can spin up the correct set
+      # of images during testing
+
+      # We don't need to build the Web Docker image since it's not yet used
+      # in the integration tests. We have a separate action to verify that it builds
+      # successfully.
+      - name: Pull Model Server Docker image
+        run: |
+          docker pull onyxdotapp/onyx-model-server:latest
+          docker tag onyxdotapp/onyx-model-server:latest onyxdotapp/onyx-model-server:test
+          
       - name: Set up Python
         uses: actions/setup-python@v5
         with:
@@ -41,6 +69,49 @@ jobs:
           pip install --retries 5 --timeout 30 -r backend/requirements/default.txt
           pip install --retries 5 --timeout 30 -r backend/requirements/dev.txt
 
+      - name: Start Docker containers
+        run: |
+          cd deployment/docker_compose
+          ENABLE_PAID_ENTERPRISE_EDITION_FEATURES=true \
+          AUTH_TYPE=basic \
+          REQUIRE_EMAIL_VERIFICATION=false \
+          DISABLE_TELEMETRY=true \
+          IMAGE_TAG=test \
+          docker compose -f docker-compose.model-server-test.yml -p onyx-stack up -d indexing_model_server
+        id: start_docker
+
+      - name: Wait for service to be ready
+        run: |
+          echo "Starting wait-for-service script..."
+
+          start_time=$(date +%s)
+          timeout=300  # 5 minutes in seconds
+
+          while true; do
+            current_time=$(date +%s)
+            elapsed_time=$((current_time - start_time))
+            
+            if [ $elapsed_time -ge $timeout ]; then
+              echo "Timeout reached. Service did not become ready in 5 minutes."
+              exit 1
+            fi
+            
+            # Use curl with error handling to ignore specific exit code 56
+            response=$(curl -s -o /dev/null -w "%{http_code}" http://localhost:9000/api/health || echo "curl_error")
+            
+            if [ "$response" = "200" ]; then
+              echo "Service is ready!"
+              break
+            elif [ "$response" = "curl_error" ]; then
+              echo "Curl encountered an error, possibly exit code 56. Continuing to retry..."
+            else
+              echo "Service not ready yet (HTTP status $response). Retrying in 5 seconds..."
+            fi
+            
+            sleep 5
+          done
+          echo "Finished waiting for service."
+          
       - name: Run Tests
         shell: script -q -e -c "bash --noprofile --norc -eo pipefail {0}"
         run: |
@@ -56,3 +127,23 @@ jobs:
             -H 'Content-type: application/json' \
             --data '{"text":"Scheduled Model Tests failed! Check the run at: https://github.com/${{ github.repository }}/actions/runs/${{ github.run_id }}"}' \
             $SLACK_WEBHOOK
+            
+      - name: Dump all-container logs (optional)
+        if: always()
+        run: |
+          cd deployment/docker_compose
+          docker compose -f docker-compose.model-server-test.yml -p onyx-stack logs --no-color > $GITHUB_WORKSPACE/docker-compose.log || true
+
+      - name: Upload logs
+        if: always()
+        uses: actions/upload-artifact@v4
+        with:
+          name: docker-all-logs
+          path: ${{ github.workspace }}/docker-compose.log
+          
+      - name: Stop Docker containers
+        if: always()
+        run: |
+          cd deployment/docker_compose
+          docker compose -f docker-compose.model-server-test.yml -p onyx-stack down -v
+          

.vscode/launch.template.jsonc

@@ -205,7 +205,7 @@
                 "--loglevel=INFO",
                 "--hostname=light@%n",
                 "-Q",
-                "vespa_metadata_sync,connector_deletion,doc_permissions_upsert",
+                "vespa_metadata_sync,connector_deletion,doc_permissions_upsert,checkpoint_cleanup",
             ],
             "presentation": {
 				 "group": "2",

README.md

@@ -24,113 +24,93 @@
 </a>
 </p>
 
-<strong>[Onyx](https://www.onyx.app/)</strong> (formerly Danswer) is the AI Assistant connected to your company's docs, apps, and people.
-Onyx provides a Chat interface and plugs into any LLM of your choice. Onyx can be deployed anywhere and for any
-scale - on a laptop, on-premise, or to cloud. Since you own the deployment, your user data and chats are fully in your
-own control. Onyx is dual Licensed with most of it under MIT license and designed to be modular and easily extensible. The system also comes fully ready
-for production usage with user authentication, role management (admin/basic users), chat persistence, and a UI for
-configuring AI Assistants.
+<strong>[Onyx](https://www.onyx.app/)</strong> (formerly Danswer) is the AI platform connected to your company's docs, apps, and people.
+Onyx provides a feature rich Chat interface and plugs into any LLM of your choice.
+Keep knowledge and access controls sync-ed across over 40 connectors like Google Drive, Slack, Confluence, Salesforce, etc.
+Create custom AI agents with unique prompts, knowledge, and actions that the agents can take.
+Onyx can be deployed securely anywhere and for any scale - on a laptop, on-premise, or to cloud.
 
-Onyx also serves as a Enterprise Search across all common workplace tools such as Slack, Google Drive, Confluence, etc.
-By combining LLMs and team specific knowledge, Onyx becomes a subject matter expert for the team. Imagine ChatGPT if
-it had access to your team's unique knowledge! It enables questions such as "A customer wants feature X, is this already
-supported?" or "Where's the pull request for feature Y?"
 
-<h3>Usage</h3>
+<h3>Feature Highlights</h3>
 
-Onyx Web App:
+**Deep research over your team's knowledge:**
 
-https://github.com/onyx-dot-app/onyx/assets/32520769/563be14c-9304-47b5-bf0a-9049c2b6f410
+https://private-user-images.githubusercontent.com/32520769/414509312-48392e83-95d0-4fb5-8650-a396e05e0a32.mp4?jwt=eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.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.a9D8A0sgKE9AoaoE-mfFbJ6_OKYeqaf7TZ4Han2JfW8
 
-Or, plug Onyx into your existing Slack workflows (more integrations to come 😁):
 
-https://github.com/onyx-dot-app/onyx/assets/25087905/3e19739b-d178-4371-9a38-011430bdec1b
+**Use Onyx as a secure AI Chat with any LLM:**
+
+![Onyx Chat Silent Demo](https://github.com/onyx-dot-app/onyx/releases/download/v0.21.1/OnyxChatSilentDemo.gif)
+
+
+**Easily set up connectors to your apps:**
+
+![Onyx Connector Silent Demo](https://github.com/onyx-dot-app/onyx/releases/download/v0.21.1/OnyxConnectorSilentDemo.gif)
+
+
+**Access Onyx where your team already works:**
+
+![Onyx Bot Demo](https://github.com/onyx-dot-app/onyx/releases/download/v0.21.1/OnyxBot.png)
 
-For more details on the Admin UI to manage connectors and users, check out our
-<strong><a href="https://www.youtube.com/watch?v=geNzY1nbCnU">Full Video Demo</a></strong>!
 
 ## Deployment
+**To try it out for free and get started in seconds, check out [Onyx Cloud](https://cloud.onyx.app/signup)**.
 
-Onyx can easily be run locally (even on a laptop) or deployed on a virtual machine with a single
+Onyx can also be run locally (even on a laptop) or deployed on a virtual machine with a single
 `docker compose` command. Checkout our [docs](https://docs.onyx.app/quickstart) to learn more.
 
-We also have built-in support for deployment on Kubernetes. Files for that can be found [here](https://github.com/onyx-dot-app/onyx/tree/main/deployment/kubernetes).
+We also have built-in support for high-availability/scalable deployment on Kubernetes.
+References [here](https://github.com/onyx-dot-app/onyx/tree/main/deployment).
 
-## 💃 Main Features
 
-- Chat UI with the ability to select documents to chat with.
-- Create custom AI Assistants with different prompts and backing knowledge sets.
-- Connect Onyx with LLM of your choice (self-host for a fully airgapped solution).
-- Document Search + AI Answers for natural language queries.
-- Connectors to all common workplace tools like Google Drive, Confluence, Slack, etc.
-- Slack integration to get answers and search results directly in Slack.
+## 🔍 Other Notable Benefits of Onyx
+- Custom deep learning models for indexing and inference time, only through Onyx + learning from user feedback.
+- Flexible security features like SSO (OIDC/SAML/OAuth2), RBAC, encryption of credentials, etc.
+- Knowledge curation features like document-sets, query history, usage analytics, etc.
+- Scalable deployment options tested up to many tens of thousands users and hundreds of millions of documents.
+
 
 ## 🚧 Roadmap
-
-- Chat/Prompt sharing with specific teammates and user groups.
-- Multimodal model support, chat with images, video etc.
-- Choosing between LLMs and parameters during chat session.
-- Tool calling and agent configurations options.
+- New methods in information retrieval (StructRAG, LightGraphRAG, etc.)
+- Personalized Search
 - Organizational understanding and ability to locate and suggest experts from your team.
+- Code Search
+- SQL and Structured Query Language
 
-## Other Notable Benefits of Onyx
-
-- User Authentication with document level access management.
-- Best in class Hybrid Search across all sources (BM-25 + prefix aware embedding models).
-- Admin Dashboard to configure connectors, document-sets, access, etc.
-- Custom deep learning models + learn from user feedback.
-- Easy deployment and ability to host Onyx anywhere of your choosing.
 
 ## 🔌 Connectors
+Keep knowledge and access up to sync across 40+ connectors:
 
-Efficiently pulls the latest changes from:
-
-- Slack
-- GitHub
 - Google Drive
 - Confluence
+- Slack
+- Gmail
+- Salesforce
+- Microsoft Sharepoint
+- Github
 - Jira
 - Zendesk
-- Gmail
-- Notion
 - Gong
-- Slab
-- Linear
-- Productboard
-- Guru
-- Bookstack
-- Document360
-- Sharepoint
-- Hubspot
+- Microsoft Teams
+- Dropbox
 - Local Files
 - Websites
 - And more ...
 
-## 📚 Editions
+See the full list [here](https://docs.onyx.app/connectors).
 
+
+## 📚 Licensing
 There are two editions of Onyx:
 
-- Onyx Community Edition (CE) is available freely under the MIT Expat license. This version has ALL the core features discussed above. This is the version of Onyx you will get if you follow the Deployment guide above.
-- Onyx Enterprise Edition (EE) includes extra features that are primarily useful for larger organizations. Specifically, this includes:
-  - Single Sign-On (SSO), with support for both SAML and OIDC
-  - Role-based access control
-  - Document permission inheritance from connected sources
-  - Usage analytics and query history accessible to admins
-  - Whitelabeling
-  - API key authentication
-  - Encryption of secrets
-  - And many more! Checkout [our website](https://www.onyx.app/) for the latest.
+- Onyx Community Edition (CE) is available freely under the MIT Expat license. Simply follow the Deployment guide above.
+- Onyx Enterprise Edition (EE) includes extra features that are primarily useful for larger organizations.
+For feature details, check out [our website](https://www.onyx.app/pricing).
 
 To try the Onyx Enterprise Edition:
+1. Checkout [Onyx Cloud](https://cloud.onyx.app/signup).
+2. For self-hosting the Enterprise Edition, contact us at [founders@onyx.app](mailto:founders@onyx.app) or book a call with us on our [Cal](https://cal.com/team/onyx/founders).
 
-1. Checkout our [Cloud product](https://cloud.onyx.app/signup).
-2. For self-hosting, contact us at [founders@onyx.app](mailto:founders@onyx.app) or book a call with us on our [Cal](https://cal.com/team/onyx/founders).
 
 ## 💡 Contributing
-
 Looking to contribute? Please check out the [Contribution Guide](CONTRIBUTING.md) for more details.
-
-## ⭐Star History
-
-[![Star History Chart](https://api.star-history.com/svg?repos=onyx-dot-app/onyx&type=Date)](https://star-history.com/#onyx-dot-app/onyx&Date)
-

backend/Dockerfile

@@ -28,11 +28,11 @@ RUN apt-get update && \
         curl \
         zip \
         ca-certificates \
-        libgnutls30=3.7.9-2+deb12u3 \
-        libblkid1=2.38.1-5+deb12u1 \
-        libmount1=2.38.1-5+deb12u1 \
-        libsmartcols1=2.38.1-5+deb12u1 \
-        libuuid1=2.38.1-5+deb12u1 \
+        libgnutls30 \
+        libblkid1 \
+        libmount1 \
+        libsmartcols1 \
+        libuuid1 \
         libxmlsec1-dev \
         pkg-config \
         gcc \

backend/alembic/versions/1a03d2c2856b_add_indexes_to_document__tag.py

@@ -0,0 +1,27 @@
+"""Add indexes to document__tag
+
+Revision ID: 1a03d2c2856b
+Revises: 9c00a2bccb83
+Create Date: 2025-02-18 10:45:13.957807
+
+"""
+from alembic import op
+
+# revision identifiers, used by Alembic.
+revision = "1a03d2c2856b"
+down_revision = "9c00a2bccb83"
+branch_labels: None = None
+depends_on: None = None
+
+
+def upgrade() -> None:
+    op.create_index(
+        op.f("ix_document__tag_tag_id"),
+        "document__tag",
+        ["tag_id"],
+        unique=False,
+    )
+
+
+def downgrade() -> None:
+    op.drop_index(op.f("ix_document__tag_tag_id"), table_name="document__tag")

backend/alembic/versions/3934b1bc7b62_update_github_connector_repo_name_to_.py

@@ -0,0 +1,125 @@
+"""Update GitHub connector repo_name to repositories
+
+Revision ID: 3934b1bc7b62
+Revises: b7c2b63c4a03
+Create Date: 2025-03-05 10:50:30.516962
+
+"""
+from alembic import op
+import sqlalchemy as sa
+import json
+import logging
+
+# revision identifiers, used by Alembic.
+revision = "3934b1bc7b62"
+down_revision = "b7c2b63c4a03"
+branch_labels = None
+depends_on = None
+
+logger = logging.getLogger("alembic.runtime.migration")
+
+
+def upgrade() -> None:
+    # Get all GitHub connectors
+    conn = op.get_bind()
+
+    # First get all GitHub connectors
+    github_connectors = conn.execute(
+        sa.text(
+            """
+            SELECT id, connector_specific_config
+            FROM connector
+            WHERE source = 'GITHUB'
+            """
+        )
+    ).fetchall()
+
+    # Update each connector's config
+    updated_count = 0
+    for connector_id, config in github_connectors:
+        try:
+            if not config:
+                logger.warning(f"Connector {connector_id} has no config, skipping")
+                continue
+
+            # Parse the config if it's a string
+            if isinstance(config, str):
+                config = json.loads(config)
+
+            if "repo_name" not in config:
+                continue
+
+            # Create new config with repositories instead of repo_name
+            new_config = dict(config)
+            repo_name_value = new_config.pop("repo_name")
+            new_config["repositories"] = repo_name_value
+
+            # Update the connector with the new config
+            conn.execute(
+                sa.text(
+                    """
+                    UPDATE connector
+                    SET connector_specific_config = :new_config
+                    WHERE id = :connector_id
+                    """
+                ),
+                {"connector_id": connector_id, "new_config": json.dumps(new_config)},
+            )
+            updated_count += 1
+        except Exception as e:
+            logger.error(f"Error updating connector {connector_id}: {str(e)}")
+
+
+def downgrade() -> None:
+    # Get all GitHub connectors
+    conn = op.get_bind()
+
+    logger.debug(
+        "Starting rollback of GitHub connectors from repositories to repo_name"
+    )
+
+    github_connectors = conn.execute(
+        sa.text(
+            """
+            SELECT id, connector_specific_config
+            FROM connector
+            WHERE source = 'GITHUB'
+            """
+        )
+    ).fetchall()
+
+    logger.debug(f"Found {len(github_connectors)} GitHub connectors to rollback")
+
+    # Revert each GitHub connector to use repo_name instead of repositories
+    reverted_count = 0
+    for connector_id, config in github_connectors:
+        try:
+            if not config:
+                continue
+
+            # Parse the config if it's a string
+            if isinstance(config, str):
+                config = json.loads(config)
+
+            if "repositories" not in config:
+                continue
+
+            # Create new config with repo_name instead of repositories
+            new_config = dict(config)
+            repositories_value = new_config.pop("repositories")
+            new_config["repo_name"] = repositories_value
+
+            # Update the connector with the new config
+            conn.execute(
+                sa.text(
+                    """
+                    UPDATE connector
+                    SET connector_specific_config = :new_config
+                    WHERE id = :connector_id
+                    """
+                ),
+                {"new_config": json.dumps(new_config), "connector_id": connector_id},
+            )
+            reverted_count += 1
+        except Exception as e:
+            logger.error(f"Error reverting connector {connector_id}: {str(e)}")

backend/alembic/versions/3bd4c84fe72f_improved_index.py

@@ -0,0 +1,84 @@
+"""improved index
+
+Revision ID: 3bd4c84fe72f
+Revises: 8f43500ee275
+Create Date: 2025-02-26 13:07:56.217791
+
+"""
+from alembic import op
+
+
+# revision identifiers, used by Alembic.
+revision = "3bd4c84fe72f"
+down_revision = "8f43500ee275"
+branch_labels = None
+depends_on = None
+
+
+# NOTE:
+# This migration addresses issues with the previous migration (8f43500ee275) which caused
+# an outage by creating an index without using CONCURRENTLY. This migration:
+#
+# 1. Creates more efficient full-text search capabilities using tsvector columns and GIN indexes
+# 2. Uses CONCURRENTLY for all index creation to prevent table locking
+# 3. Explicitly manages transactions with COMMIT statements to allow CONCURRENTLY to work
+# (see: https://www.postgresql.org/docs/9.4/sql-createindex.html#SQL-CREATEINDEX-CONCURRENTLY)
+# (see: https://github.com/sqlalchemy/alembic/issues/277)
+# 4. Adds indexes to both chat_message and chat_session tables for comprehensive search
+
+
+def upgrade() -> None:
+    # Create a GIN index for full-text search on chat_message.message
+    op.execute(
+        """
+        ALTER TABLE chat_message
+        ADD COLUMN message_tsv tsvector
+        GENERATED ALWAYS AS (to_tsvector('english', message)) STORED;
+        """
+    )
+
+    # Commit the current transaction before creating concurrent indexes
+    op.execute("COMMIT")
+
+    op.execute(
+        """
+        CREATE INDEX CONCURRENTLY IF NOT EXISTS idx_chat_message_tsv
+        ON chat_message
+        USING GIN (message_tsv)
+        """
+    )
+
+    # Also add a stored tsvector column for chat_session.description
+    op.execute(
+        """
+        ALTER TABLE chat_session
+        ADD COLUMN description_tsv tsvector
+        GENERATED ALWAYS AS (to_tsvector('english', coalesce(description, ''))) STORED;
+        """
+    )
+
+    # Commit again before creating the second concurrent index
+    op.execute("COMMIT")
+
+    op.execute(
+        """
+        CREATE INDEX CONCURRENTLY IF NOT EXISTS idx_chat_session_desc_tsv
+        ON chat_session
+        USING GIN (description_tsv)
+        """
+    )
+
+
+def downgrade() -> None:
+    # Drop the indexes first (use CONCURRENTLY for dropping too)
+    op.execute("COMMIT")
+    op.execute("DROP INDEX CONCURRENTLY IF EXISTS idx_chat_message_tsv;")
+
+    op.execute("COMMIT")
+    op.execute("DROP INDEX CONCURRENTLY IF EXISTS idx_chat_session_desc_tsv;")
+
+    # Then drop the columns
+    op.execute("ALTER TABLE chat_message DROP COLUMN IF EXISTS message_tsv;")
+    op.execute("ALTER TABLE chat_session DROP COLUMN IF EXISTS description_tsv;")
+
+    op.execute("DROP INDEX IF EXISTS idx_chat_message_message_lower;")

backend/alembic/versions/8f43500ee275_add_index.py

@@ -0,0 +1,32 @@
+"""add index
+
+Revision ID: 8f43500ee275
+Revises: da42808081e3
+Create Date: 2025-02-24 17:35:33.072714
+
+"""
+from alembic import op
+
+
+# revision identifiers, used by Alembic.
+revision = "8f43500ee275"
+down_revision = "da42808081e3"
+branch_labels = None
+depends_on = None
+
+
+def upgrade() -> None:
+    # Create a basic index on the lowercase message column for direct text matching
+    # Limit to 1500 characters to stay well under the 2856 byte limit of btree version 4
+    # op.execute(
+    #     """
+    #     CREATE INDEX idx_chat_message_message_lower
+    #     ON chat_message (LOWER(substring(message, 1, 1500)))
+    #     """
+    # )
+    pass
+
+
+def downgrade() -> None:
+    # Drop the index
+    op.execute("DROP INDEX IF EXISTS idx_chat_message_message_lower;")

backend/alembic/versions/9c00a2bccb83_chat_message_agentic.py

@@ -0,0 +1,43 @@
+"""chat_message_agentic
+
+Revision ID: 9c00a2bccb83
+Revises: b7a7eee5aa15
+Create Date: 2025-02-17 11:15:43.081150
+
+"""
+from alembic import op
+import sqlalchemy as sa
+
+
+# revision identifiers, used by Alembic.
+revision = "9c00a2bccb83"
+down_revision = "b7a7eee5aa15"
+branch_labels = None
+depends_on = None
+
+
+def upgrade() -> None:
+    # First add the column as nullable
+    op.add_column("chat_message", sa.Column("is_agentic", sa.Boolean(), nullable=True))
+
+    # Update existing rows based on presence of SubQuestions
+    op.execute(
+        """
+        UPDATE chat_message
+        SET is_agentic = EXISTS (
+            SELECT 1
+            FROM agent__sub_question
+            WHERE agent__sub_question.primary_question_id = chat_message.id
+        )
+        WHERE is_agentic IS NULL
+    """
+    )
+
+    # Make the column non-nullable with a default value of False
+    op.alter_column(
+        "chat_message", "is_agentic", nullable=False, server_default=sa.text("false")
+    )
+
+
+def downgrade() -> None:
+    op.drop_column("chat_message", "is_agentic")

backend/alembic/versions/acaab4ef4507_remove_inactive_ccpair_status_on_.py

@@ -0,0 +1,29 @@
+"""remove inactive ccpair status on downgrade
+
+Revision ID: acaab4ef4507
+Revises: b388730a2899
+Create Date: 2025-02-16 18:21:41.330212
+
+"""
+from alembic import op
+from onyx.db.models import ConnectorCredentialPair
+from onyx.db.enums import ConnectorCredentialPairStatus
+from sqlalchemy import update
+
+# revision identifiers, used by Alembic.
+revision = "acaab4ef4507"
+down_revision = "b388730a2899"
+branch_labels = None
+depends_on = None
+
+
+def upgrade() -> None:
+    pass
+
+
+def downgrade() -> None:
+    op.execute(
+        update(ConnectorCredentialPair)
+        .where(ConnectorCredentialPair.status == ConnectorCredentialPairStatus.INVALID)
+        .values(status=ConnectorCredentialPairStatus.ACTIVE)
+    )

backend/alembic/versions/b388730a2899_nullable_preferences.py

@@ -0,0 +1,31 @@
+"""nullable preferences
+
+Revision ID: b388730a2899
+Revises: 1a03d2c2856b
+Create Date: 2025-02-17 18:49:22.643902
+
+"""
+from alembic import op
+
+
+# revision identifiers, used by Alembic.
+revision = "b388730a2899"
+down_revision = "1a03d2c2856b"
+branch_labels = None
+depends_on = None
+
+
+def upgrade() -> None:
+    op.alter_column("user", "temperature_override_enabled", nullable=True)
+    op.alter_column("user", "auto_scroll", nullable=True)
+
+
+def downgrade() -> None:
+    # Ensure no null values before making columns non-nullable
+    op.execute(
+        'UPDATE "user" SET temperature_override_enabled = false WHERE temperature_override_enabled IS NULL'
+    )
+    op.execute('UPDATE "user" SET auto_scroll = false WHERE auto_scroll IS NULL')
+
+    op.alter_column("user", "temperature_override_enabled", nullable=False)
+    op.alter_column("user", "auto_scroll", nullable=False)

backend/alembic/versions/b7a7eee5aa15_add_checkpointing_failure_handling.py

@@ -0,0 +1,124 @@
+"""Add checkpointing/failure handling
+
+Revision ID: b7a7eee5aa15
+Revises: f39c5794c10a
+Create Date: 2025-01-24 15:17:36.763172
+
+"""
+from alembic import op
+import sqlalchemy as sa
+from sqlalchemy.dialects import postgresql
+
+# revision identifiers, used by Alembic.
+revision = "b7a7eee5aa15"
+down_revision = "f39c5794c10a"
+branch_labels = None
+depends_on = None
+
+
+def upgrade() -> None:
+    op.add_column(
+        "index_attempt",
+        sa.Column("checkpoint_pointer", sa.String(), nullable=True),
+    )
+    op.add_column(
+        "index_attempt",
+        sa.Column("poll_range_start", sa.DateTime(timezone=True), nullable=True),
+    )
+    op.add_column(
+        "index_attempt",
+        sa.Column("poll_range_end", sa.DateTime(timezone=True), nullable=True),
+    )
+
+    op.create_index(
+        "ix_index_attempt_cc_pair_settings_poll",
+        "index_attempt",
+        [
+            "connector_credential_pair_id",
+            "search_settings_id",
+            "status",
+            sa.text("time_updated DESC"),
+        ],
+    )
+
+    # Drop the old IndexAttemptError table
+    op.drop_index("index_attempt_id", table_name="index_attempt_errors")
+    op.drop_table("index_attempt_errors")
+
+    # Create the new version of the table
+    op.create_table(
+        "index_attempt_errors",
+        sa.Column("id", sa.Integer(), primary_key=True),
+        sa.Column("index_attempt_id", sa.Integer(), nullable=False),
+        sa.Column("connector_credential_pair_id", sa.Integer(), nullable=False),
+        sa.Column("document_id", sa.String(), nullable=True),
+        sa.Column("document_link", sa.String(), nullable=True),
+        sa.Column("entity_id", sa.String(), nullable=True),
+        sa.Column("failed_time_range_start", sa.DateTime(timezone=True), nullable=True),
+        sa.Column("failed_time_range_end", sa.DateTime(timezone=True), nullable=True),
+        sa.Column("failure_message", sa.Text(), nullable=False),
+        sa.Column("is_resolved", sa.Boolean(), nullable=False, default=False),
+        sa.Column(
+            "time_created",
+            sa.DateTime(timezone=True),
+            server_default=sa.text("now()"),
+            nullable=False,
+        ),
+        sa.ForeignKeyConstraint(
+            ["index_attempt_id"],
+            ["index_attempt.id"],
+        ),
+        sa.ForeignKeyConstraint(
+            ["connector_credential_pair_id"],
+            ["connector_credential_pair.id"],
+        ),
+    )
+
+
+def downgrade() -> None:
+    op.execute("SET lock_timeout = '5s'")
+
+    # try a few times to drop the table, this has been observed to fail due to other locks
+    # blocking the drop
+    NUM_TRIES = 10
+    for i in range(NUM_TRIES):
+        try:
+            op.drop_table("index_attempt_errors")
+            break
+        except Exception as e:
+            if i == NUM_TRIES - 1:
+                raise e
+            print(f"Error dropping table: {e}. Retrying...")
+
+    op.execute("SET lock_timeout = DEFAULT")
+
+    # Recreate the old IndexAttemptError table
+    op.create_table(
+        "index_attempt_errors",
+        sa.Column("id", sa.Integer(), primary_key=True),
+        sa.Column("index_attempt_id", sa.Integer(), nullable=True),
+        sa.Column("batch", sa.Integer(), nullable=True),
+        sa.Column("doc_summaries", postgresql.JSONB(), nullable=False),
+        sa.Column("error_msg", sa.Text(), nullable=True),
+        sa.Column("traceback", sa.Text(), nullable=True),
+        sa.Column(
+            "time_created",
+            sa.DateTime(timezone=True),
+            server_default=sa.text("now()"),
+        ),
+        sa.ForeignKeyConstraint(
+            ["index_attempt_id"],
+            ["index_attempt.id"],
+        ),
+    )
+
+    op.create_index(
+        "index_attempt_id",
+        "index_attempt_errors",
+        ["time_created"],
+    )
+
+    op.drop_index("ix_index_attempt_cc_pair_settings_poll")
+    op.drop_column("index_attempt", "checkpoint_pointer")
+    op.drop_column("index_attempt", "poll_range_start")
+    op.drop_column("index_attempt", "poll_range_end")

backend/alembic/versions/b7c2b63c4a03_add_background_reindex_enabled_field.py

@@ -0,0 +1,55 @@
+"""add background_reindex_enabled field
+
+Revision ID: b7c2b63c4a03
+Revises: f11b408e39d3
+Create Date: 2024-03-26 12:34:56.789012
+
+"""
+from alembic import op
+import sqlalchemy as sa
+
+from onyx.db.enums import EmbeddingPrecision
+
+
+# revision identifiers, used by Alembic.
+revision = "b7c2b63c4a03"
+down_revision = "f11b408e39d3"
+branch_labels = None
+depends_on = None
+
+
+def upgrade() -> None:
+    # Add background_reindex_enabled column with default value of True
+    op.add_column(
+        "search_settings",
+        sa.Column(
+            "background_reindex_enabled",
+            sa.Boolean(),
+            nullable=False,
+            server_default="true",
+        ),
+    )
+
+    # Add embedding_precision column with default value of FLOAT
+    op.add_column(
+        "search_settings",
+        sa.Column(
+            "embedding_precision",
+            sa.Enum(EmbeddingPrecision, native_enum=False),
+            nullable=False,
+            server_default=EmbeddingPrecision.FLOAT.name,
+        ),
+    )
+
+    # Add reduced_dimension column with default value of None
+    op.add_column(
+        "search_settings",
+        sa.Column("reduced_dimension", sa.Integer(), nullable=True),
+    )
+
+
+def downgrade() -> None:
+    # Remove the background_reindex_enabled column
+    op.drop_column("search_settings", "background_reindex_enabled")
+    op.drop_column("search_settings", "embedding_precision")
+    op.drop_column("search_settings", "reduced_dimension")

backend/alembic/versions/da42808081e3_migrate_jira_connectors_to_new_format.py

@@ -0,0 +1,120 @@
+"""migrate jira connectors to new format
+
+Revision ID: da42808081e3
+Revises: f13db29f3101
+Create Date: 2025-02-24 11:24:54.396040
+
+"""
+from alembic import op
+import sqlalchemy as sa
+import json
+
+from onyx.configs.constants import DocumentSource
+from onyx.connectors.onyx_jira.utils import extract_jira_project
+
+
+# revision identifiers, used by Alembic.
+revision = "da42808081e3"
+down_revision = "f13db29f3101"
+branch_labels = None
+depends_on = None
+
+
+def upgrade() -> None:
+    # Get all Jira connectors
+    conn = op.get_bind()
+
+    # First get all Jira connectors
+    jira_connectors = conn.execute(
+        sa.text(
+            """
+            SELECT id, connector_specific_config
+            FROM connector
+            WHERE source = :source
+            """
+        ),
+        {"source": DocumentSource.JIRA.value.upper()},
+    ).fetchall()
+
+    # Update each connector's config
+    for connector_id, old_config in jira_connectors:
+        if not old_config:
+            continue
+
+        # Extract project key from URL if it exists
+        new_config: dict[str, str | None] = {}
+        if project_url := old_config.get("jira_project_url"):
+            # Parse the URL to get base and project
+            try:
+                jira_base, project_key = extract_jira_project(project_url)
+                new_config = {"jira_base_url": jira_base, "project_key": project_key}
+            except ValueError:
+                # If URL parsing fails, just use the URL as the base
+                new_config = {
+                    "jira_base_url": project_url.split("/projects/")[0],
+                    "project_key": None,
+                }
+        else:
+            # For connectors without a project URL, we need admin intervention
+            # Mark these for review
+            print(
+                f"WARNING: Jira connector {connector_id} has no project URL configured"
+            )
+            continue
+
+        # Update the connector config
+        conn.execute(
+            sa.text(
+                """
+                UPDATE connector
+                SET connector_specific_config = :new_config
+                WHERE id = :id
+                """
+            ),
+            {"id": connector_id, "new_config": json.dumps(new_config)},
+        )
+
+
+def downgrade() -> None:
+    # Get all Jira connectors
+    conn = op.get_bind()
+
+    # First get all Jira connectors
+    jira_connectors = conn.execute(
+        sa.text(
+            """
+            SELECT id, connector_specific_config
+            FROM connector
+            WHERE source = :source
+            """
+        ),
+        {"source": DocumentSource.JIRA.value.upper()},
+    ).fetchall()
+
+    # Update each connector's config back to the old format
+    for connector_id, new_config in jira_connectors:
+        if not new_config:
+            continue
+
+        old_config = {}
+        base_url = new_config.get("jira_base_url")
+        project_key = new_config.get("project_key")
+
+        if base_url and project_key:
+            old_config = {"jira_project_url": f"{base_url}/projects/{project_key}"}
+        elif base_url:
+            old_config = {"jira_project_url": base_url}
+        else:
+            continue
+
+        # Update the connector config
+        conn.execute(
+            sa.text(
+                """
+                UPDATE connector
+                SET connector_specific_config = :old_config
+                WHERE id = :id
+                """
+            ),
+            {"id": connector_id, "old_config": old_config},
+        )

backend/alembic/versions/f11b408e39d3_force_lowercase_all_users.py

@@ -0,0 +1,36 @@
+"""force lowercase all users
+
+Revision ID: f11b408e39d3
+Revises: 3bd4c84fe72f
+Create Date: 2025-02-26 17:04:55.683500
+
+"""
+
+
+# revision identifiers, used by Alembic.
+revision = "f11b408e39d3"
+down_revision = "3bd4c84fe72f"
+branch_labels = None
+depends_on = None
+
+
+def upgrade() -> None:
+    # 1) Convert all existing user emails to lowercase
+    from alembic import op
+
+    op.execute(
+        """
+        UPDATE "user"
+        SET email = LOWER(email)
+        """
+    )
+
+    # 2) Add a check constraint to ensure emails are always lowercase
+    op.create_check_constraint("ensure_lowercase_email", "user", "email = LOWER(email)")
+
+
+def downgrade() -> None:
+    # Drop the check constraint
+    from alembic import op
+
+    op.drop_constraint("ensure_lowercase_email", "user", type_="check")

backend/alembic/versions/f13db29f3101_add_composite_index_for_last_modified_.py

@@ -0,0 +1,27 @@
+"""Add composite index for last_modified and last_synced to document
+
+Revision ID: f13db29f3101
+Revises: b388730a2899
+Create Date: 2025-02-18 22:48:11.511389
+
+"""
+from alembic import op
+
+# revision identifiers, used by Alembic.
+revision = "f13db29f3101"
+down_revision = "acaab4ef4507"
+branch_labels: str | None = None
+depends_on: str | None = None
+
+
+def upgrade() -> None:
+    op.create_index(
+        "ix_document_sync_status",
+        "document",
+        ["last_modified", "last_synced"],
+        unique=False,
+    )
+
+
+def downgrade() -> None:
+    op.drop_index("ix_document_sync_status", table_name="document")

backend/alembic_tenants/versions/34e3630c7f32_lowercase_multi_tenant_user_auth.py

@@ -0,0 +1,42 @@
+"""lowercase multi-tenant user auth
+
+Revision ID: 34e3630c7f32
+Revises: a4f6ee863c47
+Create Date: 2025-02-26 15:03:01.211894
+
+"""
+from alembic import op
+
+
+# revision identifiers, used by Alembic.
+revision = "34e3630c7f32"
+down_revision = "a4f6ee863c47"
+branch_labels = None
+depends_on = None
+
+
+def upgrade() -> None:
+    # 1) Convert all existing rows to lowercase
+    op.execute(
+        """
+        UPDATE user_tenant_mapping
+        SET email = LOWER(email)
+        """
+    )
+    # 2) Add a check constraint so that emails cannot be written in uppercase
+    op.create_check_constraint(
+        "ensure_lowercase_email",
+        "user_tenant_mapping",
+        "email = LOWER(email)",
+        schema="public",
+    )
+
+
+def downgrade() -> None:
+    # Drop the check constraint
+    op.drop_constraint(
+        "ensure_lowercase_email",
+        "user_tenant_mapping",
+        schema="public",
+        type_="check",
+    )

backend/ee/onyx/background/celery/apps/primary.py

@@ -4,12 +4,11 @@ from ee.onyx.server.reporting.usage_export_generation import create_new_usage_re
 from onyx.background.celery.apps.primary import celery_app
 from onyx.background.task_utils import build_celery_task_wrapper
 from onyx.configs.app_configs import JOB_TIMEOUT
-from onyx.db.chat import delete_chat_sessions_older_than
-from onyx.db.engine import get_session_with_tenant
+from onyx.db.chat import delete_chat_session
+from onyx.db.chat import get_chat_sessions_older_than
+from onyx.db.engine import get_session_with_current_tenant
 from onyx.server.settings.store import load_settings
 from onyx.utils.logger import setup_logger
-from shared_configs.configs import MULTI_TENANT
-from shared_configs.contextvars import CURRENT_TENANT_ID_CONTEXTVAR
 
 logger = setup_logger()
 
@@ -18,11 +17,28 @@ logger = setup_logger()
 
 @build_celery_task_wrapper(name_chat_ttl_task)
 @celery_app.task(soft_time_limit=JOB_TIMEOUT)
-def perform_ttl_management_task(
-    retention_limit_days: int, *, tenant_id: str | None
-) -> None:
-    with get_session_with_tenant(tenant_id) as db_session:
-        delete_chat_sessions_older_than(retention_limit_days, db_session)
+def perform_ttl_management_task(retention_limit_days: int, *, tenant_id: str) -> None:
+    with get_session_with_current_tenant() as db_session:
+        old_chat_sessions = get_chat_sessions_older_than(
+            retention_limit_days, db_session
+        )
+
+    for user_id, session_id in old_chat_sessions:
+        # one session per delete so that we don't blow up if a deletion fails.
+        with get_session_with_current_tenant() as db_session:
+            try:
+                delete_chat_session(
+                    user_id,
+                    session_id,
+                    db_session,
+                    include_deleted=True,
+                    hard_delete=True,
+                )
+            except Exception:
+                logger.exception(
+                    "delete_chat_session exceptioned. "
+                    f"user_id={user_id} session_id={session_id}"
+                )
 
 
 #####
@@ -35,24 +51,19 @@ def perform_ttl_management_task(
     ignore_result=True,
     soft_time_limit=JOB_TIMEOUT,
 )
-def check_ttl_management_task(*, tenant_id: str | None) -> None:
+def check_ttl_management_task(*, tenant_id: str) -> None:
     """Runs periodically to check if any ttl tasks should be run and adds them
     to the queue"""
-    token = None
-    if MULTI_TENANT and tenant_id is not None:
-        token = CURRENT_TENANT_ID_CONTEXTVAR.set(tenant_id)
 
     settings = load_settings()
     retention_limit_days = settings.maximum_chat_retention_days
-    with get_session_with_tenant(tenant_id) as db_session:
+    with get_session_with_current_tenant() as db_session:
         if should_perform_chat_ttl_check(retention_limit_days, db_session):
             perform_ttl_management_task.apply_async(
                 kwargs=dict(
                     retention_limit_days=retention_limit_days, tenant_id=tenant_id
                 ),
             )
-    if token is not None:
-        CURRENT_TENANT_ID_CONTEXTVAR.reset(token)
 
 
 @celery_app.task(
@@ -60,9 +71,9 @@ def check_ttl_management_task(*, tenant_id: str | None) -> None:
     ignore_result=True,
     soft_time_limit=JOB_TIMEOUT,
 )
-def autogenerate_usage_report_task(*, tenant_id: str | None) -> None:
+def autogenerate_usage_report_task(*, tenant_id: str) -> None:
     """This generates usage report under the /admin/generate-usage/report endpoint"""
-    with get_session_with_tenant(tenant_id) as db_session:
+    with get_session_with_current_tenant() as db_session:
         create_new_usage_report(
             db_session=db_session,
             user_id=None,

backend/ee/onyx/background/celery/tasks/vespa/tasks.py

@@ -18,7 +18,7 @@ logger = setup_logger()
 
 
 def monitor_usergroup_taskset(
-    tenant_id: str | None, key_bytes: bytes, r: Redis, db_session: Session
+    tenant_id: str, key_bytes: bytes, r: Redis, db_session: Session
 ) -> None:
     """This function is likely to move in the worker refactor happening next."""
     fence_key = key_bytes.decode("utf-8")

backend/ee/onyx/configs/app_configs.py

@@ -59,10 +59,14 @@ SUPER_CLOUD_API_KEY = os.environ.get("SUPER_CLOUD_API_KEY", "api_key")
 
 OAUTH_SLACK_CLIENT_ID = os.environ.get("OAUTH_SLACK_CLIENT_ID", "")
 OAUTH_SLACK_CLIENT_SECRET = os.environ.get("OAUTH_SLACK_CLIENT_SECRET", "")
-OAUTH_CONFLUENCE_CLIENT_ID = os.environ.get("OAUTH_CONFLUENCE_CLIENT_ID", "")
-OAUTH_CONFLUENCE_CLIENT_SECRET = os.environ.get("OAUTH_CONFLUENCE_CLIENT_SECRET", "")
-OAUTH_JIRA_CLIENT_ID = os.environ.get("OAUTH_JIRA_CLIENT_ID", "")
-OAUTH_JIRA_CLIENT_SECRET = os.environ.get("OAUTH_JIRA_CLIENT_SECRET", "")
+OAUTH_CONFLUENCE_CLOUD_CLIENT_ID = os.environ.get(
+    "OAUTH_CONFLUENCE_CLOUD_CLIENT_ID", ""
+)
+OAUTH_CONFLUENCE_CLOUD_CLIENT_SECRET = os.environ.get(
+    "OAUTH_CONFLUENCE_CLOUD_CLIENT_SECRET", ""
+)
+OAUTH_JIRA_CLOUD_CLIENT_ID = os.environ.get("OAUTH_JIRA_CLOUD_CLIENT_ID", "")
+OAUTH_JIRA_CLOUD_CLIENT_SECRET = os.environ.get("OAUTH_JIRA_CLOUD_CLIENT_SECRET", "")
 OAUTH_GOOGLE_DRIVE_CLIENT_ID = os.environ.get("OAUTH_GOOGLE_DRIVE_CLIENT_ID", "")
 OAUTH_GOOGLE_DRIVE_CLIENT_SECRET = os.environ.get(
     "OAUTH_GOOGLE_DRIVE_CLIENT_SECRET", ""

backend/ee/onyx/db/connector_credential_pair.py

@@ -4,6 +4,7 @@ from sqlalchemy.orm import Session
 from onyx.configs.constants import DocumentSource
 from onyx.db.connector_credential_pair import get_connector_credential_pair
 from onyx.db.enums import AccessType
+from onyx.db.enums import ConnectorCredentialPairStatus
 from onyx.db.models import Connector
 from onyx.db.models import ConnectorCredentialPair
 from onyx.db.models import UserGroup__ConnectorCredentialPair
@@ -35,10 +36,11 @@ def _delete_connector_credential_pair_user_groups_relationship__no_commit(
 def get_cc_pairs_by_source(
     db_session: Session,
     source_type: DocumentSource,
-    only_sync: bool,
+    access_type: AccessType | None = None,
+    status: ConnectorCredentialPairStatus | None = None,
 ) -> list[ConnectorCredentialPair]:
     """
-    Get all cc_pairs for a given source type (and optionally only sync)
+    Get all cc_pairs for a given source type with optional filtering by access_type and status
     result is sorted by cc_pair id
     """
     query = (
@@ -48,8 +50,11 @@ def get_cc_pairs_by_source(
         .order_by(ConnectorCredentialPair.id)
     )
 
-    if only_sync:
-        query = query.filter(ConnectorCredentialPair.access_type == AccessType.SYNC)
+    if access_type is not None:
+        query = query.filter(ConnectorCredentialPair.access_type == access_type)
+
+    if status is not None:
+        query = query.filter(ConnectorCredentialPair.status == status)
 
     cc_pairs = query.all()
     return cc_pairs

backend/ee/onyx/db/query_history.py

@@ -134,7 +134,9 @@ def fetch_chat_sessions_eagerly_by_time(
     limit: int | None = 500,
     initial_time: datetime | None = None,
 ) -> list[ChatSession]:
-    time_order: UnaryExpression = desc(ChatSession.time_created)
+    """Sorted by oldest to newest, then by message id"""
+
+    asc_time_order: UnaryExpression = asc(ChatSession.time_created)
     message_order: UnaryExpression = asc(ChatMessage.id)
 
     filters: list[ColumnElement | BinaryExpression] = [
@@ -147,8 +149,7 @@ def fetch_chat_sessions_eagerly_by_time(
     subquery = (
         db_session.query(ChatSession.id, ChatSession.time_created)
         .filter(*filters)
-        .order_by(ChatSession.id, time_order)
-        .distinct(ChatSession.id)
+        .order_by(asc_time_order)
         .limit(limit)
         .subquery()
     )
@@ -164,7 +165,7 @@ def fetch_chat_sessions_eagerly_by_time(
                 ChatMessage.chat_message_feedbacks
             ),
         )
-        .order_by(time_order, message_order)
+        .order_by(asc_time_order, message_order)
     )
 
     chat_sessions = query.all()

backend/ee/onyx/db/usage_export.py

@@ -16,13 +16,18 @@ from onyx.db.models import UsageReport
 from onyx.file_store.file_store import get_default_file_store
 
 
-# Gets skeletons of all message
+# Gets skeletons of all messages in the given range
 def get_empty_chat_messages_entries__paginated(
     db_session: Session,
     period: tuple[datetime, datetime],
     limit: int | None = 500,
     initial_time: datetime | None = None,
 ) -> tuple[Optional[datetime], list[ChatMessageSkeleton]]:
+    """Returns a tuple where:
+    first element is the most recent timestamp out of the sessions iterated
+    - this timestamp can be used to paginate forward in time
+    second element is a list of messages belonging to all the sessions iterated
+    """
     chat_sessions = fetch_chat_sessions_eagerly_by_time(
         start=period[0],
         end=period[1],
@@ -52,18 +57,17 @@ def get_empty_chat_messages_entries__paginated(
     if len(chat_sessions) == 0:
         return None, []
 
-    return chat_sessions[0].time_created, message_skeletons
+    return chat_sessions[-1].time_created, message_skeletons
 
 
 def get_all_empty_chat_message_entries(
     db_session: Session,
     period: tuple[datetime, datetime],
 ) -> Generator[list[ChatMessageSkeleton], None, None]:
+    """period is the range of time over which to fetch messages."""
     initial_time: Optional[datetime] = period[0]
-    ind = 0
     while True:
-        ind += 1
-
+        # iterate from oldest to newest
         time_created, message_skeletons = get_empty_chat_messages_entries__paginated(
             db_session,
             period,

backend/ee/onyx/db/user_group.py

@@ -424,7 +424,7 @@ def _validate_curator_status__no_commit(
         )
 
         # if the user is a curator in any of their groups, set their role to CURATOR
-        # otherwise, set their role to BASIC
+        # otherwise, set their role to BASIC only if they were previously a CURATOR
         if curator_relationships:
             user.role = UserRole.CURATOR
         elif user.role == UserRole.CURATOR:
@@ -631,7 +631,16 @@ def update_user_group(
     removed_users = db_session.scalars(
         select(User).where(User.id.in_(removed_user_ids))  # type: ignore
     ).unique()
-    _validate_curator_status__no_commit(db_session, list(removed_users))
+
+    # Filter out admin and global curator users before validating curator status
+    users_to_validate = [
+        user
+        for user in removed_users
+        if user.role not in [UserRole.ADMIN, UserRole.GLOBAL_CURATOR]
+    ]
+
+    if users_to_validate:
+        _validate_curator_status__no_commit(db_session, users_to_validate)
 
     # update "time_updated" to now
     db_user_group.time_last_modified_by_user = func.now()

backend/ee/onyx/external_permissions/confluence/doc_sync.py

@@ -9,12 +9,16 @@ from ee.onyx.external_permissions.confluence.constants import ALL_CONF_EMAILS_GR
 from onyx.access.models import DocExternalAccess
 from onyx.access.models import ExternalAccess
 from onyx.connectors.confluence.connector import ConfluenceConnector
+from onyx.connectors.confluence.onyx_confluence import (
+    get_user_email_from_username__server,
+)
 from onyx.connectors.confluence.onyx_confluence import OnyxConfluence
-from onyx.connectors.confluence.utils import get_user_email_from_username__server
+from onyx.connectors.credentials_provider import OnyxDBCredentialsProvider
 from onyx.connectors.models import SlimDocument
 from onyx.db.models import ConnectorCredentialPair
 from onyx.indexing.indexing_heartbeat import IndexingHeartbeatInterface
 from onyx.utils.logger import setup_logger
+from shared_configs.contextvars import get_current_tenant_id
 
 logger = setup_logger()
 
@@ -342,7 +346,8 @@ def _fetch_all_page_restrictions(
 
 
 def confluence_doc_sync(
-    cc_pair: ConnectorCredentialPair, callback: IndexingHeartbeatInterface | None
+    cc_pair: ConnectorCredentialPair,
+    callback: IndexingHeartbeatInterface | None,
 ) -> list[DocExternalAccess]:
     """
     Adds the external permissions to the documents in postgres
@@ -354,7 +359,11 @@ def confluence_doc_sync(
     confluence_connector = ConfluenceConnector(
         **cc_pair.connector.connector_specific_config
     )
-    confluence_connector.load_credentials(cc_pair.credential.credential_json)
+
+    provider = OnyxDBCredentialsProvider(
+        get_current_tenant_id(), "confluence", cc_pair.credential_id
+    )
+    confluence_connector.set_credentials_provider(provider)
 
     is_cloud = cc_pair.connector.connector_specific_config.get("is_cloud", False)
 

backend/ee/onyx/external_permissions/confluence/group_sync.py

@@ -1,9 +1,11 @@
 from ee.onyx.db.external_perm import ExternalUserGroup
 from ee.onyx.external_permissions.confluence.constants import ALL_CONF_EMAILS_GROUP_NAME
 from onyx.background.error_logging import emit_background_error
-from onyx.connectors.confluence.onyx_confluence import build_confluence_client
+from onyx.connectors.confluence.onyx_confluence import (
+    get_user_email_from_username__server,
+)
 from onyx.connectors.confluence.onyx_confluence import OnyxConfluence
-from onyx.connectors.confluence.utils import get_user_email_from_username__server
+from onyx.connectors.credentials_provider import OnyxDBCredentialsProvider
 from onyx.db.models import ConnectorCredentialPair
 from onyx.utils.logger import setup_logger
 
@@ -14,30 +16,24 @@ def _build_group_member_email_map(
     confluence_client: OnyxConfluence, cc_pair_id: int
 ) -> dict[str, set[str]]:
     group_member_emails: dict[str, set[str]] = {}
-    for user_result in confluence_client.paginated_cql_user_retrieval():
-        logger.debug(f"Processing groups for user: {user_result}")
+    for user in confluence_client.paginated_cql_user_retrieval():
+        logger.debug(f"Processing groups for user: {user}")
 
-        user = user_result.get("user", {})
-        if not user:
-            msg = f"user result missing user field: {user_result}"
-            emit_background_error(msg, cc_pair_id=cc_pair_id)
-            logger.error(msg)
-            continue
-
-        email = user.get("email")
+        email = user.email
         if not email:
             # This field is only present in Confluence Server
-            user_name = user.get("username")
+            user_name = user.username
             # If it is present, try to get the email using a Server-specific method
             if user_name:
                 email = get_user_email_from_username__server(
                     confluence_client=confluence_client,
                     user_name=user_name,
                 )
+
         if not email:
             # If we still don't have an email, skip this user
-            msg = f"user result missing email field: {user_result}"
-            if user.get("type") == "app":
+            msg = f"user result missing email field: {user}"
+            if user.type == "app":
                 logger.warning(msg)
             else:
                 emit_background_error(msg, cc_pair_id=cc_pair_id)
@@ -45,7 +41,7 @@ def _build_group_member_email_map(
             continue
 
         all_users_groups: set[str] = set()
-        for group in confluence_client.paginated_groups_by_user_retrieval(user):
+        for group in confluence_client.paginated_groups_by_user_retrieval(user.user_id):
             # group name uniqueness is enforced by Confluence, so we can use it as a group ID
             group_id = group["name"]
             group_member_emails.setdefault(group_id, set()).add(email)
@@ -67,13 +63,27 @@ def _build_group_member_email_map(
 
 
 def confluence_group_sync(
+    tenant_id: str,
     cc_pair: ConnectorCredentialPair,
 ) -> list[ExternalUserGroup]:
-    confluence_client = build_confluence_client(
-        credentials=cc_pair.credential.credential_json,
-        is_cloud=cc_pair.connector.connector_specific_config.get("is_cloud", False),
-        wiki_base=cc_pair.connector.connector_specific_config["wiki_base"],
-    )
+    provider = OnyxDBCredentialsProvider(tenant_id, "confluence", cc_pair.credential_id)
+    is_cloud = cc_pair.connector.connector_specific_config.get("is_cloud", False)
+    wiki_base: str = cc_pair.connector.connector_specific_config["wiki_base"]
+    url = wiki_base.rstrip("/")
+
+    probe_kwargs = {
+        "max_backoff_retries": 6,
+        "max_backoff_seconds": 10,
+    }
+
+    final_kwargs = {
+        "max_backoff_retries": 10,
+        "max_backoff_seconds": 60,
+    }
+
+    confluence_client = OnyxConfluence(is_cloud, url, provider)
+    confluence_client._probe_connection(**probe_kwargs)
+    confluence_client._initialize_connection(**final_kwargs)
 
     group_member_email_map = _build_group_member_email_map(
         confluence_client=confluence_client,

backend/ee/onyx/external_permissions/gmail/doc_sync.py

@@ -32,7 +32,8 @@ def _get_slim_doc_generator(
 
 
 def gmail_doc_sync(
-    cc_pair: ConnectorCredentialPair, callback: IndexingHeartbeatInterface | None
+    cc_pair: ConnectorCredentialPair,
+    callback: IndexingHeartbeatInterface | None,
 ) -> list[DocExternalAccess]:
     """
     Adds the external permissions to the documents in postgres

backend/ee/onyx/external_permissions/google_drive/doc_sync.py

@@ -62,12 +62,14 @@ def _fetch_permissions_for_permission_ids(
         user_email=(owner_email or google_drive_connector.primary_admin_email),
     )
 
+    # We continue on 404 or 403 because the document may not exist or the user may not have access to it
     fetched_permissions = execute_paginated_retrieval(
         retrieval_function=drive_service.permissions().list,
         list_key="permissions",
         fileId=doc_id,
         fields="permissions(id, emailAddress, type, domain)",
         supportsAllDrives=True,
+        continue_on_404_or_403=True,
     )
 
     permissions_for_doc_id = []
@@ -104,7 +106,13 @@ def _get_permissions_from_slim_doc(
     user_emails: set[str] = set()
     group_emails: set[str] = set()
     public = False
+    skipped_permissions = 0
+
     for permission in permissions_list:
+        if not permission:
+            skipped_permissions += 1
+            continue
+
         permission_type = permission["type"]
         if permission_type == "user":
             user_emails.add(permission["emailAddress"])
@@ -121,6 +129,11 @@ def _get_permissions_from_slim_doc(
         elif permission_type == "anyone":
             public = True
 
+    if skipped_permissions > 0:
+        logger.warning(
+            f"Skipped {skipped_permissions} permissions of {len(permissions_list)} for document {slim_doc.id}"
+        )
+
     drive_id = permission_info.get("drive_id")
     group_ids = group_emails | ({drive_id} if drive_id is not None else set())
 
@@ -132,7 +145,8 @@ def _get_permissions_from_slim_doc(
 
 
 def gdrive_doc_sync(
-    cc_pair: ConnectorCredentialPair, callback: IndexingHeartbeatInterface | None
+    cc_pair: ConnectorCredentialPair,
+    callback: IndexingHeartbeatInterface | None,
 ) -> list[DocExternalAccess]:
     """
     Adds the external permissions to the documents in postgres

backend/ee/onyx/external_permissions/google_drive/group_sync.py

@@ -119,6 +119,7 @@ def _build_onyx_groups(
 
 
 def gdrive_group_sync(
+    tenant_id: str,
     cc_pair: ConnectorCredentialPair,
 ) -> list[ExternalUserGroup]:
     # Initialize connector and build credential/service objects

backend/ee/onyx/external_permissions/slack/doc_sync.py

@@ -5,7 +5,7 @@ from onyx.access.models import DocExternalAccess
 from onyx.access.models import ExternalAccess
 from onyx.connectors.slack.connector import get_channels
 from onyx.connectors.slack.connector import make_paginated_slack_api_call_w_retries
-from onyx.connectors.slack.connector import SlackPollConnector
+from onyx.connectors.slack.connector import SlackConnector
 from onyx.db.models import ConnectorCredentialPair
 from onyx.indexing.indexing_heartbeat import IndexingHeartbeatInterface
 from onyx.utils.logger import setup_logger
@@ -17,7 +17,7 @@ logger = setup_logger()
 def _get_slack_document_ids_and_channels(
     cc_pair: ConnectorCredentialPair, callback: IndexingHeartbeatInterface | None
 ) -> dict[str, list[str]]:
-    slack_connector = SlackPollConnector(**cc_pair.connector.connector_specific_config)
+    slack_connector = SlackConnector(**cc_pair.connector.connector_specific_config)
     slack_connector.load_credentials(cc_pair.credential.credential_json)
 
     slim_doc_generator = slack_connector.retrieve_all_slim_documents(callback=callback)
@@ -123,7 +123,8 @@ def _fetch_channel_permissions(
 
 
 def slack_doc_sync(
-    cc_pair: ConnectorCredentialPair, callback: IndexingHeartbeatInterface | None
+    cc_pair: ConnectorCredentialPair,
+    callback: IndexingHeartbeatInterface | None,
 ) -> list[DocExternalAccess]:
     """
     Adds the external permissions to the documents in postgres

backend/ee/onyx/external_permissions/sync_params.py

@@ -28,6 +28,7 @@ DocSyncFuncType = Callable[
 
 GroupSyncFuncType = Callable[
     [
+        str,
         ConnectorCredentialPair,
     ],
     list[ExternalUserGroup],

backend/ee/onyx/main.py

@@ -15,7 +15,7 @@ from ee.onyx.server.enterprise_settings.api import (
 )
 from ee.onyx.server.manage.standard_answer import router as standard_answer_router
 from ee.onyx.server.middleware.tenant_tracking import add_tenant_id_middleware
-from ee.onyx.server.oauth import router as oauth_router
+from ee.onyx.server.oauth.api import router as ee_oauth_router
 from ee.onyx.server.query_and_chat.chat_backend import (
     router as chat_router,
 )
@@ -128,7 +128,7 @@ def get_application() -> FastAPI:
     include_router_with_global_prefix_prepended(application, query_router)
     include_router_with_global_prefix_prepended(application, chat_router)
     include_router_with_global_prefix_prepended(application, standard_answer_router)
-    include_router_with_global_prefix_prepended(application, oauth_router)
+    include_router_with_global_prefix_prepended(application, ee_oauth_router)
 
     # Enterprise-only global settings
     include_router_with_global_prefix_prepended(
@@ -152,4 +152,8 @@ def get_application() -> FastAPI:
     # environment variable. Used to automate deployment for multiple environments.
     seed_db()
 
+    # for debugging discovered routes
+    # for route in application.router.routes:
+    #     print(f"Path: {route.path}, Methods: {route.methods}")
+
     return application

backend/ee/onyx/onyxbot/slack/handlers/handle_standard_answers.py

@@ -22,7 +22,7 @@ from onyx.onyxbot.slack.blocks import get_restate_blocks
 from onyx.onyxbot.slack.constants import GENERATE_ANSWER_BUTTON_ACTION_ID
 from onyx.onyxbot.slack.handlers.utils import send_team_member_message
 from onyx.onyxbot.slack.models import SlackMessageInfo
-from onyx.onyxbot.slack.utils import respond_in_thread
+from onyx.onyxbot.slack.utils import respond_in_thread_or_channel
 from onyx.onyxbot.slack.utils import update_emote_react
 from onyx.utils.logger import OnyxLoggingAdapter
 from onyx.utils.logger import setup_logger
@@ -216,7 +216,7 @@ def _handle_standard_answers(
         all_blocks = restate_question_blocks + answer_blocks
 
         try:
-            respond_in_thread(
+            respond_in_thread_or_channel(
                 client=client,
                 channel=message_info.channel_to_respond,
                 receiver_ids=receiver_ids,
@@ -231,6 +231,7 @@ def _handle_standard_answers(
                     client=client,
                     channel=message_info.channel_to_respond,
                     thread_ts=slack_thread_id,
+                    receiver_ids=receiver_ids,
                 )
 
             return True

backend/ee/onyx/server/middleware/tenant_tracking.py

@@ -33,7 +33,7 @@ def add_tenant_id_middleware(app: FastAPI, logger: logging.LoggerAdapter) -> Non
             return await call_next(request)
 
         except Exception as e:
-            logger.error(f"Error in tenant ID middleware: {str(e)}")
+            logger.exception(f"Error in tenant ID middleware: {str(e)}")
             raise
 
 
@@ -49,7 +49,7 @@ async def _get_tenant_id_from_request(
     """
     # Check for API key
     tenant_id = extract_tenant_from_api_key_header(request)
-    if tenant_id:
+    if tenant_id is not None:
         return tenant_id
 
     # Check for anonymous user cookie

backend/ee/onyx/server/oauth.py

@@ -1,631 +0,0 @@
-import base64
-import json
-import uuid
-from typing import Any
-from typing import cast
-
-import requests
-from fastapi import APIRouter
-from fastapi import Depends
-from fastapi import HTTPException
-from fastapi.responses import JSONResponse
-from pydantic import BaseModel
-from sqlalchemy.orm import Session
-
-from ee.onyx.configs.app_configs import OAUTH_CONFLUENCE_CLIENT_ID
-from ee.onyx.configs.app_configs import OAUTH_CONFLUENCE_CLIENT_SECRET
-from ee.onyx.configs.app_configs import OAUTH_GOOGLE_DRIVE_CLIENT_ID
-from ee.onyx.configs.app_configs import OAUTH_GOOGLE_DRIVE_CLIENT_SECRET
-from ee.onyx.configs.app_configs import OAUTH_SLACK_CLIENT_ID
-from ee.onyx.configs.app_configs import OAUTH_SLACK_CLIENT_SECRET
-from onyx.auth.users import current_user
-from onyx.configs.app_configs import WEB_DOMAIN
-from onyx.configs.constants import DocumentSource
-from onyx.connectors.google_utils.google_auth import get_google_oauth_creds
-from onyx.connectors.google_utils.google_auth import sanitize_oauth_credentials
-from onyx.connectors.google_utils.shared_constants import (
-    DB_CREDENTIALS_AUTHENTICATION_METHOD,
-)
-from onyx.connectors.google_utils.shared_constants import (
-    DB_CREDENTIALS_DICT_TOKEN_KEY,
-)
-from onyx.connectors.google_utils.shared_constants import (
-    DB_CREDENTIALS_PRIMARY_ADMIN_KEY,
-)
-from onyx.connectors.google_utils.shared_constants import (
-    GoogleOAuthAuthenticationMethod,
-)
-from onyx.db.credentials import create_credential
-from onyx.db.engine import get_current_tenant_id
-from onyx.db.engine import get_session
-from onyx.db.models import User
-from onyx.redis.redis_pool import get_redis_client
-from onyx.server.documents.models import CredentialBase
-from onyx.utils.logger import setup_logger
-
-
-logger = setup_logger()
-
-router = APIRouter(prefix="/oauth")
-
-
-class SlackOAuth:
-    # https://knock.app/blog/how-to-authenticate-users-in-slack-using-oauth
-    # Example: https://api.slack.com/authentication/oauth-v2#exchanging
-
-    class OAuthSession(BaseModel):
-        """Stored in redis to be looked up on callback"""
-
-        email: str
-        redirect_on_success: str | None  # Where to send the user if OAuth flow succeeds
-
-    CLIENT_ID = OAUTH_SLACK_CLIENT_ID
-    CLIENT_SECRET = OAUTH_SLACK_CLIENT_SECRET
-
-    TOKEN_URL = "https://slack.com/api/oauth.v2.access"
-
-    # SCOPE is per https://docs.onyx.app/connectors/slack
-    BOT_SCOPE = (
-        "channels:history,"
-        "channels:read,"
-        "groups:history,"
-        "groups:read,"
-        "channels:join,"
-        "im:history,"
-        "users:read,"
-        "users:read.email,"
-        "usergroups:read"
-    )
-
-    REDIRECT_URI = f"{WEB_DOMAIN}/admin/connectors/slack/oauth/callback"
-    DEV_REDIRECT_URI = f"https://redirectmeto.com/{REDIRECT_URI}"
-
-    @classmethod
-    def generate_oauth_url(cls, state: str) -> str:
-        return cls._generate_oauth_url_helper(cls.REDIRECT_URI, state)
-
-    @classmethod
-    def generate_dev_oauth_url(cls, state: str) -> str:
-        """dev mode workaround for localhost testing
-        - https://www.nango.dev/blog/oauth-redirects-on-localhost-with-https
-        """
-
-        return cls._generate_oauth_url_helper(cls.DEV_REDIRECT_URI, state)
-
-    @classmethod
-    def _generate_oauth_url_helper(cls, redirect_uri: str, state: str) -> str:
-        url = (
-            f"https://slack.com/oauth/v2/authorize"
-            f"?client_id={cls.CLIENT_ID}"
-            f"&redirect_uri={redirect_uri}"
-            f"&scope={cls.BOT_SCOPE}"
-            f"&state={state}"
-        )
-        return url
-
-    @classmethod
-    def session_dump_json(cls, email: str, redirect_on_success: str | None) -> str:
-        """Temporary state to store in redis. to be looked up on auth response.
-        Returns a json string.
-        """
-        session = SlackOAuth.OAuthSession(
-            email=email, redirect_on_success=redirect_on_success
-        )
-        return session.model_dump_json()
-
-    @classmethod
-    def parse_session(cls, session_json: str) -> OAuthSession:
-        session = SlackOAuth.OAuthSession.model_validate_json(session_json)
-        return session
-
-
-class ConfluenceCloudOAuth:
-    """work in progress"""
-
-    # https://developer.atlassian.com/cloud/confluence/oauth-2-3lo-apps/
-
-    class OAuthSession(BaseModel):
-        """Stored in redis to be looked up on callback"""
-
-        email: str
-        redirect_on_success: str | None  # Where to send the user if OAuth flow succeeds
-
-    CLIENT_ID = OAUTH_CONFLUENCE_CLIENT_ID
-    CLIENT_SECRET = OAUTH_CONFLUENCE_CLIENT_SECRET
-    TOKEN_URL = "https://auth.atlassian.com/oauth/token"
-
-    # All read scopes per https://developer.atlassian.com/cloud/confluence/scopes-for-oauth-2-3LO-and-forge-apps/
-    CONFLUENCE_OAUTH_SCOPE = (
-        "read:confluence-props%20"
-        "read:confluence-content.all%20"
-        "read:confluence-content.summary%20"
-        "read:confluence-content.permission%20"
-        "read:confluence-user%20"
-        "read:confluence-groups%20"
-        "readonly:content.attachment:confluence"
-    )
-
-    REDIRECT_URI = f"{WEB_DOMAIN}/admin/connectors/confluence/oauth/callback"
-    DEV_REDIRECT_URI = f"https://redirectmeto.com/{REDIRECT_URI}"
-
-    # eventually for Confluence Data Center
-    # oauth_url = (
-    #     f"http://localhost:8090/rest/oauth/v2/authorize?client_id={CONFLUENCE_OAUTH_CLIENT_ID}"
-    #     f"&scope={CONFLUENCE_OAUTH_SCOPE_2}"
-    #     f"&redirect_uri={redirectme_uri}"
-    # )
-
-    @classmethod
-    def generate_oauth_url(cls, state: str) -> str:
-        return cls._generate_oauth_url_helper(cls.REDIRECT_URI, state)
-
-    @classmethod
-    def generate_dev_oauth_url(cls, state: str) -> str:
-        """dev mode workaround for localhost testing
-        - https://www.nango.dev/blog/oauth-redirects-on-localhost-with-https
-        """
-        return cls._generate_oauth_url_helper(cls.DEV_REDIRECT_URI, state)
-
-    @classmethod
-    def _generate_oauth_url_helper(cls, redirect_uri: str, state: str) -> str:
-        url = (
-            "https://auth.atlassian.com/authorize"
-            f"?audience=api.atlassian.com"
-            f"&client_id={cls.CLIENT_ID}"
-            f"&redirect_uri={redirect_uri}"
-            f"&scope={cls.CONFLUENCE_OAUTH_SCOPE}"
-            f"&state={state}"
-            "&response_type=code"
-            "&prompt=consent"
-        )
-        return url
-
-    @classmethod
-    def session_dump_json(cls, email: str, redirect_on_success: str | None) -> str:
-        """Temporary state to store in redis. to be looked up on auth response.
-        Returns a json string.
-        """
-        session = ConfluenceCloudOAuth.OAuthSession(
-            email=email, redirect_on_success=redirect_on_success
-        )
-        return session.model_dump_json()
-
-    @classmethod
-    def parse_session(cls, session_json: str) -> SlackOAuth.OAuthSession:
-        session = SlackOAuth.OAuthSession.model_validate_json(session_json)
-        return session
-
-
-class GoogleDriveOAuth:
-    # https://developers.google.com/identity/protocols/oauth2
-    # https://developers.google.com/identity/protocols/oauth2/web-server
-
-    class OAuthSession(BaseModel):
-        """Stored in redis to be looked up on callback"""
-
-        email: str
-        redirect_on_success: str | None  # Where to send the user if OAuth flow succeeds
-
-    CLIENT_ID = OAUTH_GOOGLE_DRIVE_CLIENT_ID
-    CLIENT_SECRET = OAUTH_GOOGLE_DRIVE_CLIENT_SECRET
-
-    TOKEN_URL = "https://oauth2.googleapis.com/token"
-
-    # SCOPE is per https://docs.onyx.app/connectors/google-drive
-    # TODO: Merge with or use google_utils.GOOGLE_SCOPES
-    SCOPE = (
-        "https://www.googleapis.com/auth/drive.readonly%20"
-        "https://www.googleapis.com/auth/drive.metadata.readonly%20"
-        "https://www.googleapis.com/auth/admin.directory.user.readonly%20"
-        "https://www.googleapis.com/auth/admin.directory.group.readonly"
-    )
-
-    REDIRECT_URI = f"{WEB_DOMAIN}/admin/connectors/google-drive/oauth/callback"
-    DEV_REDIRECT_URI = f"https://redirectmeto.com/{REDIRECT_URI}"
-
-    @classmethod
-    def generate_oauth_url(cls, state: str) -> str:
-        return cls._generate_oauth_url_helper(cls.REDIRECT_URI, state)
-
-    @classmethod
-    def generate_dev_oauth_url(cls, state: str) -> str:
-        """dev mode workaround for localhost testing
-        - https://www.nango.dev/blog/oauth-redirects-on-localhost-with-https
-        """
-
-        return cls._generate_oauth_url_helper(cls.DEV_REDIRECT_URI, state)
-
-    @classmethod
-    def _generate_oauth_url_helper(cls, redirect_uri: str, state: str) -> str:
-        # without prompt=consent, a refresh token is only issued the first time the user approves
-        url = (
-            f"https://accounts.google.com/o/oauth2/v2/auth"
-            f"?client_id={cls.CLIENT_ID}"
-            f"&redirect_uri={redirect_uri}"
-            "&response_type=code"
-            f"&scope={cls.SCOPE}"
-            "&access_type=offline"
-            f"&state={state}"
-            "&prompt=consent"
-        )
-        return url
-
-    @classmethod
-    def session_dump_json(cls, email: str, redirect_on_success: str | None) -> str:
-        """Temporary state to store in redis. to be looked up on auth response.
-        Returns a json string.
-        """
-        session = GoogleDriveOAuth.OAuthSession(
-            email=email, redirect_on_success=redirect_on_success
-        )
-        return session.model_dump_json()
-
-    @classmethod
-    def parse_session(cls, session_json: str) -> OAuthSession:
-        session = GoogleDriveOAuth.OAuthSession.model_validate_json(session_json)
-        return session
-
-
-@router.post("/prepare-authorization-request")
-def prepare_authorization_request(
-    connector: DocumentSource,
-    redirect_on_success: str | None,
-    user: User = Depends(current_user),
-    tenant_id: str | None = Depends(get_current_tenant_id),
-) -> JSONResponse:
-    """Used by the frontend to generate the url for the user's browser during auth request.
-
-    Example: https://www.oauth.com/oauth2-servers/authorization/the-authorization-request/
-    """
-
-    # create random oauth state param for security and to retrieve user data later
-    oauth_uuid = uuid.uuid4()
-    oauth_uuid_str = str(oauth_uuid)
-
-    # urlsafe b64 encode the uuid for the oauth url
-    oauth_state = (
-        base64.urlsafe_b64encode(oauth_uuid.bytes).rstrip(b"=").decode("utf-8")
-    )
-    session: str
-
-    if connector == DocumentSource.SLACK:
-        oauth_url = SlackOAuth.generate_oauth_url(oauth_state)
-        session = SlackOAuth.session_dump_json(
-            email=user.email, redirect_on_success=redirect_on_success
-        )
-    elif connector == DocumentSource.GOOGLE_DRIVE:
-        oauth_url = GoogleDriveOAuth.generate_oauth_url(oauth_state)
-        session = GoogleDriveOAuth.session_dump_json(
-            email=user.email, redirect_on_success=redirect_on_success
-        )
-    # elif connector == DocumentSource.CONFLUENCE:
-    #     oauth_url = ConfluenceCloudOAuth.generate_oauth_url(oauth_state)
-    #     session = ConfluenceCloudOAuth.session_dump_json(
-    #         email=user.email, redirect_on_success=redirect_on_success
-    #     )
-    # elif connector == DocumentSource.JIRA:
-    #     oauth_url = JiraCloudOAuth.generate_dev_oauth_url(oauth_state)
-    else:
-        oauth_url = None
-
-    if not oauth_url:
-        raise HTTPException(
-            status_code=404,
-            detail=f"The document source type {connector} does not have OAuth implemented",
-        )
-
-    r = get_redis_client(tenant_id=tenant_id)
-
-    # store important session state to retrieve when the user is redirected back
-    # 10 min is the max we want an oauth flow to be valid
-    r.set(f"da_oauth:{oauth_uuid_str}", session, ex=600)
-
-    return JSONResponse(content={"url": oauth_url})
-
-
-@router.post("/connector/slack/callback")
-def handle_slack_oauth_callback(
-    code: str,
-    state: str,
-    user: User = Depends(current_user),
-    db_session: Session = Depends(get_session),
-    tenant_id: str | None = Depends(get_current_tenant_id),
-) -> JSONResponse:
-    if not SlackOAuth.CLIENT_ID or not SlackOAuth.CLIENT_SECRET:
-        raise HTTPException(
-            status_code=500,
-            detail="Slack client ID or client secret is not configured.",
-        )
-
-    r = get_redis_client(tenant_id=tenant_id)
-
-    # recover the state
-    padded_state = state + "=" * (
-        -len(state) % 4
-    )  # Add padding back (Base64 decoding requires padding)
-    uuid_bytes = base64.urlsafe_b64decode(
-        padded_state
-    )  # Decode the Base64 string back to bytes
-
-    # Convert bytes back to a UUID
-    oauth_uuid = uuid.UUID(bytes=uuid_bytes)
-    oauth_uuid_str = str(oauth_uuid)
-
-    r_key = f"da_oauth:{oauth_uuid_str}"
-
-    session_json_bytes = cast(bytes, r.get(r_key))
-    if not session_json_bytes:
-        raise HTTPException(
-            status_code=400,
-            detail=f"Slack OAuth failed - OAuth state key not found: key={r_key}",
-        )
-
-    session_json = session_json_bytes.decode("utf-8")
-    try:
-        session = SlackOAuth.parse_session(session_json)
-
-        # Exchange the authorization code for an access token
-        response = requests.post(
-            SlackOAuth.TOKEN_URL,
-            headers={"Content-Type": "application/x-www-form-urlencoded"},
-            data={
-                "client_id": SlackOAuth.CLIENT_ID,
-                "client_secret": SlackOAuth.CLIENT_SECRET,
-                "code": code,
-                "redirect_uri": SlackOAuth.REDIRECT_URI,
-            },
-        )
-
-        response_data = response.json()
-
-        if not response_data.get("ok"):
-            raise HTTPException(
-                status_code=400,
-                detail=f"Slack OAuth failed: {response_data.get('error')}",
-            )
-
-        # Extract token and team information
-        access_token: str = response_data.get("access_token")
-        team_id: str = response_data.get("team", {}).get("id")
-        authed_user_id: str = response_data.get("authed_user", {}).get("id")
-
-        credential_info = CredentialBase(
-            credential_json={"slack_bot_token": access_token},
-            admin_public=True,
-            source=DocumentSource.SLACK,
-            name="Slack OAuth",
-        )
-
-        create_credential(credential_info, user, db_session)
-    except Exception as e:
-        return JSONResponse(
-            status_code=500,
-            content={
-                "success": False,
-                "message": f"An error occurred during Slack OAuth: {str(e)}",
-            },
-        )
-    finally:
-        r.delete(r_key)
-
-    # return the result
-    return JSONResponse(
-        content={
-            "success": True,
-            "message": "Slack OAuth completed successfully.",
-            "team_id": team_id,
-            "authed_user_id": authed_user_id,
-            "redirect_on_success": session.redirect_on_success,
-        }
-    )
-
-
-# Work in progress
-# @router.post("/connector/confluence/callback")
-# def handle_confluence_oauth_callback(
-#     code: str,
-#     state: str,
-#     user: User = Depends(current_user),
-#     db_session: Session = Depends(get_session),
-#     tenant_id: str | None = Depends(get_current_tenant_id),
-# ) -> JSONResponse:
-#     if not ConfluenceCloudOAuth.CLIENT_ID or not ConfluenceCloudOAuth.CLIENT_SECRET:
-#         raise HTTPException(
-#             status_code=500,
-#             detail="Confluence client ID or client secret is not configured."
-#         )
-
-#     r = get_redis_client(tenant_id=tenant_id)
-
-#     # recover the state
-#     padded_state = state + '=' * (-len(state) % 4)  # Add padding back (Base64 decoding requires padding)
-#     uuid_bytes = base64.urlsafe_b64decode(padded_state)  # Decode the Base64 string back to bytes
-
-#     # Convert bytes back to a UUID
-#     oauth_uuid = uuid.UUID(bytes=uuid_bytes)
-#     oauth_uuid_str = str(oauth_uuid)
-
-#     r_key = f"da_oauth:{oauth_uuid_str}"
-
-#     result = r.get(r_key)
-#     if not result:
-#         raise HTTPException(
-#             status_code=400,
-#             detail=f"Confluence OAuth failed - OAuth state key not found: key={r_key}"
-#         )
-
-#     try:
-#         session = ConfluenceCloudOAuth.parse_session(result)
-
-#         # Exchange the authorization code for an access token
-#         response = requests.post(
-#             ConfluenceCloudOAuth.TOKEN_URL,
-#             headers={"Content-Type": "application/x-www-form-urlencoded"},
-#             data={
-#                 "client_id": ConfluenceCloudOAuth.CLIENT_ID,
-#                 "client_secret": ConfluenceCloudOAuth.CLIENT_SECRET,
-#                 "code": code,
-#                 "redirect_uri": ConfluenceCloudOAuth.DEV_REDIRECT_URI,
-#             },
-#         )
-
-#         response_data = response.json()
-
-#         if not response_data.get("ok"):
-#             raise HTTPException(
-#                 status_code=400,
-#                 detail=f"ConfluenceCloudOAuth OAuth failed: {response_data.get('error')}"
-#             )
-
-#         # Extract token and team information
-#         access_token: str = response_data.get("access_token")
-#         team_id: str = response_data.get("team", {}).get("id")
-#         authed_user_id: str = response_data.get("authed_user", {}).get("id")
-
-#         credential_info = CredentialBase(
-#             credential_json={"slack_bot_token": access_token},
-#             admin_public=True,
-#             source=DocumentSource.CONFLUENCE,
-#             name="Confluence OAuth",
-#         )
-
-#         logger.info(f"Slack access token: {access_token}")
-
-#         credential = create_credential(credential_info, user, db_session)
-
-#         logger.info(f"new_credential_id={credential.id}")
-#     except Exception as e:
-#         return JSONResponse(
-#             status_code=500,
-#             content={
-#                 "success": False,
-#                 "message": f"An error occurred during Slack OAuth: {str(e)}",
-#             },
-#         )
-#     finally:
-#         r.delete(r_key)
-
-#     # return the result
-#     return JSONResponse(
-#         content={
-#             "success": True,
-#             "message": "Slack OAuth completed successfully.",
-#             "team_id": team_id,
-#             "authed_user_id": authed_user_id,
-#             "redirect_on_success": session.redirect_on_success,
-#         }
-#     )
-
-
-@router.post("/connector/google-drive/callback")
-def handle_google_drive_oauth_callback(
-    code: str,
-    state: str,
-    user: User = Depends(current_user),
-    db_session: Session = Depends(get_session),
-    tenant_id: str | None = Depends(get_current_tenant_id),
-) -> JSONResponse:
-    if not GoogleDriveOAuth.CLIENT_ID or not GoogleDriveOAuth.CLIENT_SECRET:
-        raise HTTPException(
-            status_code=500,
-            detail="Google Drive client ID or client secret is not configured.",
-        )
-
-    r = get_redis_client(tenant_id=tenant_id)
-
-    # recover the state
-    padded_state = state + "=" * (
-        -len(state) % 4
-    )  # Add padding back (Base64 decoding requires padding)
-    uuid_bytes = base64.urlsafe_b64decode(
-        padded_state
-    )  # Decode the Base64 string back to bytes
-
-    # Convert bytes back to a UUID
-    oauth_uuid = uuid.UUID(bytes=uuid_bytes)
-    oauth_uuid_str = str(oauth_uuid)
-
-    r_key = f"da_oauth:{oauth_uuid_str}"
-
-    session_json_bytes = cast(bytes, r.get(r_key))
-    if not session_json_bytes:
-        raise HTTPException(
-            status_code=400,
-            detail=f"Google Drive OAuth failed - OAuth state key not found: key={r_key}",
-        )
-
-    session_json = session_json_bytes.decode("utf-8")
-    session: GoogleDriveOAuth.OAuthSession
-    try:
-        session = GoogleDriveOAuth.parse_session(session_json)
-
-        # Exchange the authorization code for an access token
-        response = requests.post(
-            GoogleDriveOAuth.TOKEN_URL,
-            headers={"Content-Type": "application/x-www-form-urlencoded"},
-            data={
-                "client_id": GoogleDriveOAuth.CLIENT_ID,
-                "client_secret": GoogleDriveOAuth.CLIENT_SECRET,
-                "code": code,
-                "redirect_uri": GoogleDriveOAuth.REDIRECT_URI,
-                "grant_type": "authorization_code",
-            },
-        )
-
-        response.raise_for_status()
-
-        authorization_response: dict[str, Any] = response.json()
-
-        # the connector wants us to store the json in its authorized_user_info format
-        # returned from OAuthCredentials.get_authorized_user_info().
-        # So refresh immediately via get_google_oauth_creds with the params filled in
-        # from fields in authorization_response to get the json we need
-        authorized_user_info = {}
-        authorized_user_info["client_id"] = OAUTH_GOOGLE_DRIVE_CLIENT_ID
-        authorized_user_info["client_secret"] = OAUTH_GOOGLE_DRIVE_CLIENT_SECRET
-        authorized_user_info["refresh_token"] = authorization_response["refresh_token"]
-
-        token_json_str = json.dumps(authorized_user_info)
-        oauth_creds = get_google_oauth_creds(
-            token_json_str=token_json_str, source=DocumentSource.GOOGLE_DRIVE
-        )
-        if not oauth_creds:
-            raise RuntimeError("get_google_oauth_creds returned None.")
-
-        # save off the credentials
-        oauth_creds_sanitized_json_str = sanitize_oauth_credentials(oauth_creds)
-
-        credential_dict: dict[str, str] = {}
-        credential_dict[DB_CREDENTIALS_DICT_TOKEN_KEY] = oauth_creds_sanitized_json_str
-        credential_dict[DB_CREDENTIALS_PRIMARY_ADMIN_KEY] = session.email
-        credential_dict[
-            DB_CREDENTIALS_AUTHENTICATION_METHOD
-        ] = GoogleOAuthAuthenticationMethod.OAUTH_INTERACTIVE.value
-
-        credential_info = CredentialBase(
-            credential_json=credential_dict,
-            admin_public=True,
-            source=DocumentSource.GOOGLE_DRIVE,
-            name="OAuth (interactive)",
-        )
-
-        create_credential(credential_info, user, db_session)
-    except Exception as e:
-        return JSONResponse(
-            status_code=500,
-            content={
-                "success": False,
-                "message": f"An error occurred during Google Drive OAuth: {str(e)}",
-            },
-        )
-    finally:
-        r.delete(r_key)
-
-    # return the result
-    return JSONResponse(
-        content={
-            "success": True,
-            "message": "Google Drive OAuth completed successfully.",
-            "redirect_on_success": session.redirect_on_success,
-        }
-    )

backend/ee/onyx/server/oauth/api.py

@@ -0,0 +1,91 @@
+import base64
+import uuid
+
+from fastapi import Depends
+from fastapi import HTTPException
+from fastapi.responses import JSONResponse
+
+from ee.onyx.server.oauth.api_router import router
+from ee.onyx.server.oauth.confluence_cloud import ConfluenceCloudOAuth
+from ee.onyx.server.oauth.google_drive import GoogleDriveOAuth
+from ee.onyx.server.oauth.slack import SlackOAuth
+from onyx.auth.users import current_admin_user
+from onyx.configs.app_configs import DEV_MODE
+from onyx.configs.constants import DocumentSource
+from onyx.db.engine import get_current_tenant_id
+from onyx.db.models import User
+from onyx.redis.redis_pool import get_redis_client
+from onyx.utils.logger import setup_logger
+
+logger = setup_logger()
+
+
+@router.post("/prepare-authorization-request")
+def prepare_authorization_request(
+    connector: DocumentSource,
+    redirect_on_success: str | None,
+    user: User = Depends(current_admin_user),
+    tenant_id: str | None = Depends(get_current_tenant_id),
+) -> JSONResponse:
+    """Used by the frontend to generate the url for the user's browser during auth request.
+
+    Example: https://www.oauth.com/oauth2-servers/authorization/the-authorization-request/
+    """
+
+    # create random oauth state param for security and to retrieve user data later
+    oauth_uuid = uuid.uuid4()
+    oauth_uuid_str = str(oauth_uuid)
+
+    # urlsafe b64 encode the uuid for the oauth url
+    oauth_state = (
+        base64.urlsafe_b64encode(oauth_uuid.bytes).rstrip(b"=").decode("utf-8")
+    )
+
+    session: str | None = None
+    if connector == DocumentSource.SLACK:
+        if not DEV_MODE:
+            oauth_url = SlackOAuth.generate_oauth_url(oauth_state)
+        else:
+            oauth_url = SlackOAuth.generate_dev_oauth_url(oauth_state)
+
+        session = SlackOAuth.session_dump_json(
+            email=user.email, redirect_on_success=redirect_on_success
+        )
+    elif connector == DocumentSource.CONFLUENCE:
+        if not DEV_MODE:
+            oauth_url = ConfluenceCloudOAuth.generate_oauth_url(oauth_state)
+        else:
+            oauth_url = ConfluenceCloudOAuth.generate_dev_oauth_url(oauth_state)
+        session = ConfluenceCloudOAuth.session_dump_json(
+            email=user.email, redirect_on_success=redirect_on_success
+        )
+    elif connector == DocumentSource.GOOGLE_DRIVE:
+        if not DEV_MODE:
+            oauth_url = GoogleDriveOAuth.generate_oauth_url(oauth_state)
+        else:
+            oauth_url = GoogleDriveOAuth.generate_dev_oauth_url(oauth_state)
+        session = GoogleDriveOAuth.session_dump_json(
+            email=user.email, redirect_on_success=redirect_on_success
+        )
+    else:
+        oauth_url = None
+
+    if not oauth_url:
+        raise HTTPException(
+            status_code=404,
+            detail=f"The document source type {connector} does not have OAuth implemented",
+        )
+
+    if not session:
+        raise HTTPException(
+            status_code=500,
+            detail=f"The document source type {connector} failed to generate an OAuth session.",
+        )
+
+    r = get_redis_client(tenant_id=tenant_id)
+
+    # store important session state to retrieve when the user is redirected back
+    # 10 min is the max we want an oauth flow to be valid
+    r.set(f"da_oauth:{oauth_uuid_str}", session, ex=600)
+
+    return JSONResponse(content={"url": oauth_url})

backend/ee/onyx/server/oauth/api_router.py

@@ -0,0 +1,3 @@
+from fastapi import APIRouter
+
+router: APIRouter = APIRouter(prefix="/oauth")

backend/ee/onyx/server/oauth/confluence_cloud.py

@@ -0,0 +1,362 @@
+import base64
+import uuid
+from datetime import datetime
+from datetime import timedelta
+from datetime import timezone
+from typing import Any
+from typing import cast
+
+import requests
+from fastapi import Depends
+from fastapi import HTTPException
+from fastapi.responses import JSONResponse
+from pydantic import BaseModel
+from pydantic import ValidationError
+from sqlalchemy.orm import Session
+
+from ee.onyx.configs.app_configs import OAUTH_CONFLUENCE_CLOUD_CLIENT_ID
+from ee.onyx.configs.app_configs import OAUTH_CONFLUENCE_CLOUD_CLIENT_SECRET
+from ee.onyx.server.oauth.api_router import router
+from onyx.auth.users import current_admin_user
+from onyx.configs.app_configs import DEV_MODE
+from onyx.configs.app_configs import WEB_DOMAIN
+from onyx.configs.constants import DocumentSource
+from onyx.connectors.confluence.utils import CONFLUENCE_OAUTH_TOKEN_URL
+from onyx.db.credentials import create_credential
+from onyx.db.credentials import fetch_credential_by_id_for_user
+from onyx.db.credentials import update_credential_json
+from onyx.db.engine import get_current_tenant_id
+from onyx.db.engine import get_session
+from onyx.db.models import User
+from onyx.redis.redis_pool import get_redis_client
+from onyx.server.documents.models import CredentialBase
+from onyx.utils.logger import setup_logger
+
+logger = setup_logger()
+
+
+class ConfluenceCloudOAuth:
+    # https://developer.atlassian.com/cloud/confluence/oauth-2-3lo-apps/
+
+    class OAuthSession(BaseModel):
+        """Stored in redis to be looked up on callback"""
+
+        email: str
+        redirect_on_success: str | None  # Where to send the user if OAuth flow succeeds
+
+    class TokenResponse(BaseModel):
+        access_token: str
+        expires_in: int
+        token_type: str
+        refresh_token: str
+        scope: str
+
+    class AccessibleResources(BaseModel):
+        id: str
+        name: str
+        url: str
+        scopes: list[str]
+        avatarUrl: str
+
+    CLIENT_ID = OAUTH_CONFLUENCE_CLOUD_CLIENT_ID
+    CLIENT_SECRET = OAUTH_CONFLUENCE_CLOUD_CLIENT_SECRET
+    TOKEN_URL = CONFLUENCE_OAUTH_TOKEN_URL
+
+    ACCESSIBLE_RESOURCE_URL = (
+        "https://api.atlassian.com/oauth/token/accessible-resources"
+    )
+
+    # All read scopes per https://developer.atlassian.com/cloud/confluence/scopes-for-oauth-2-3LO-and-forge-apps/
+    CONFLUENCE_OAUTH_SCOPE = (
+        # classic scope
+        "read:confluence-space.summary%20"
+        "read:confluence-props%20"
+        "read:confluence-content.all%20"
+        "read:confluence-content.summary%20"
+        "read:confluence-content.permission%20"
+        "read:confluence-user%20"
+        "read:confluence-groups%20"
+        "readonly:content.attachment:confluence%20"
+        "search:confluence%20"
+        # granular scope
+        "read:attachment:confluence%20"  # possibly unneeded unless calling v2 attachments api
+        "read:content-details:confluence%20"  # for permission sync
+        "offline_access"
+    )
+
+    REDIRECT_URI = f"{WEB_DOMAIN}/admin/connectors/confluence/oauth/callback"
+    DEV_REDIRECT_URI = f"https://redirectmeto.com/{REDIRECT_URI}"
+
+    # eventually for Confluence Data Center
+    # oauth_url = (
+    #     f"http://localhost:8090/rest/oauth/v2/authorize?client_id={CONFLUENCE_OAUTH_CLIENT_ID}"
+    #     f"&scope={CONFLUENCE_OAUTH_SCOPE_2}"
+    #     f"&redirect_uri={redirectme_uri}"
+    # )
+
+    @classmethod
+    def generate_oauth_url(cls, state: str) -> str:
+        return cls._generate_oauth_url_helper(cls.REDIRECT_URI, state)
+
+    @classmethod
+    def generate_dev_oauth_url(cls, state: str) -> str:
+        """dev mode workaround for localhost testing
+        - https://www.nango.dev/blog/oauth-redirects-on-localhost-with-https
+        """
+        return cls._generate_oauth_url_helper(cls.DEV_REDIRECT_URI, state)
+
+    @classmethod
+    def _generate_oauth_url_helper(cls, redirect_uri: str, state: str) -> str:
+        # https://developer.atlassian.com/cloud/jira/platform/oauth-2-3lo-apps/#1--direct-the-user-to-the-authorization-url-to-get-an-authorization-code
+
+        url = (
+            "https://auth.atlassian.com/authorize"
+            f"?audience=api.atlassian.com"
+            f"&client_id={cls.CLIENT_ID}"
+            f"&scope={cls.CONFLUENCE_OAUTH_SCOPE}"
+            f"&redirect_uri={redirect_uri}"
+            f"&state={state}"
+            "&response_type=code"
+            "&prompt=consent"
+        )
+        return url
+
+    @classmethod
+    def session_dump_json(cls, email: str, redirect_on_success: str | None) -> str:
+        """Temporary state to store in redis. to be looked up on auth response.
+        Returns a json string.
+        """
+        session = ConfluenceCloudOAuth.OAuthSession(
+            email=email, redirect_on_success=redirect_on_success
+        )
+        return session.model_dump_json()
+
+    @classmethod
+    def parse_session(cls, session_json: str) -> OAuthSession:
+        session = ConfluenceCloudOAuth.OAuthSession.model_validate_json(session_json)
+        return session
+
+    @classmethod
+    def generate_finalize_url(cls, credential_id: int) -> str:
+        return f"{WEB_DOMAIN}/admin/connectors/confluence/oauth/finalize?credential={credential_id}"
+
+
+@router.post("/connector/confluence/callback")
+def confluence_oauth_callback(
+    code: str,
+    state: str,
+    user: User = Depends(current_admin_user),
+    db_session: Session = Depends(get_session),
+    tenant_id: str | None = Depends(get_current_tenant_id),
+) -> JSONResponse:
+    """Handles the backend logic for the frontend page that the user is redirected to
+    after visiting the oauth authorization url."""
+
+    if not ConfluenceCloudOAuth.CLIENT_ID or not ConfluenceCloudOAuth.CLIENT_SECRET:
+        raise HTTPException(
+            status_code=500,
+            detail="Confluence Cloud client ID or client secret is not configured.",
+        )
+
+    r = get_redis_client(tenant_id=tenant_id)
+
+    # recover the state
+    padded_state = state + "=" * (
+        -len(state) % 4
+    )  # Add padding back (Base64 decoding requires padding)
+    uuid_bytes = base64.urlsafe_b64decode(
+        padded_state
+    )  # Decode the Base64 string back to bytes
+
+    # Convert bytes back to a UUID
+    oauth_uuid = uuid.UUID(bytes=uuid_bytes)
+    oauth_uuid_str = str(oauth_uuid)
+
+    r_key = f"da_oauth:{oauth_uuid_str}"
+
+    session_json_bytes = cast(bytes, r.get(r_key))
+    if not session_json_bytes:
+        raise HTTPException(
+            status_code=400,
+            detail=f"Confluence Cloud OAuth failed - OAuth state key not found: key={r_key}",
+        )
+
+    session_json = session_json_bytes.decode("utf-8")
+    try:
+        session = ConfluenceCloudOAuth.parse_session(session_json)
+
+        if not DEV_MODE:
+            redirect_uri = ConfluenceCloudOAuth.REDIRECT_URI
+        else:
+            redirect_uri = ConfluenceCloudOAuth.DEV_REDIRECT_URI
+
+        # Exchange the authorization code for an access token
+        response = requests.post(
+            ConfluenceCloudOAuth.TOKEN_URL,
+            headers={"Content-Type": "application/x-www-form-urlencoded"},
+            data={
+                "client_id": ConfluenceCloudOAuth.CLIENT_ID,
+                "client_secret": ConfluenceCloudOAuth.CLIENT_SECRET,
+                "code": code,
+                "redirect_uri": redirect_uri,
+                "grant_type": "authorization_code",
+            },
+        )
+
+        token_response: ConfluenceCloudOAuth.TokenResponse | None = None
+
+        try:
+            token_response = ConfluenceCloudOAuth.TokenResponse.model_validate_json(
+                response.text
+            )
+        except Exception:
+            raise RuntimeError(
+                "Confluence Cloud OAuth failed during code/token exchange."
+            )
+
+        now = datetime.now(timezone.utc)
+        expires_at = now + timedelta(seconds=token_response.expires_in)
+
+        credential_info = CredentialBase(
+            credential_json={
+                "confluence_access_token": token_response.access_token,
+                "confluence_refresh_token": token_response.refresh_token,
+                "created_at": now.isoformat(),
+                "expires_at": expires_at.isoformat(),
+                "expires_in": token_response.expires_in,
+                "scope": token_response.scope,
+            },
+            admin_public=True,
+            source=DocumentSource.CONFLUENCE,
+            name="Confluence Cloud OAuth",
+        )
+
+        credential = create_credential(credential_info, user, db_session)
+    except Exception as e:
+        return JSONResponse(
+            status_code=500,
+            content={
+                "success": False,
+                "message": f"An error occurred during Confluence Cloud OAuth: {str(e)}",
+            },
+        )
+    finally:
+        r.delete(r_key)
+
+    # return the result
+    return JSONResponse(
+        content={
+            "success": True,
+            "message": "Confluence Cloud OAuth completed successfully.",
+            "finalize_url": ConfluenceCloudOAuth.generate_finalize_url(credential.id),
+            "redirect_on_success": session.redirect_on_success,
+        }
+    )
+
+
+@router.get("/connector/confluence/accessible-resources")
+def confluence_oauth_accessible_resources(
+    credential_id: int,
+    user: User = Depends(current_admin_user),
+    db_session: Session = Depends(get_session),
+    tenant_id: str | None = Depends(get_current_tenant_id),
+) -> JSONResponse:
+    """Atlassian's API is weird and does not supply us with enough info to be in a
+    usable state after authorizing.  All API's require a cloud id. We have to list
+    the accessible resources/sites and let the user choose which site to use."""
+
+    credential = fetch_credential_by_id_for_user(credential_id, user, db_session)
+    if not credential:
+        raise HTTPException(400, f"Credential {credential_id} not found.")
+
+    credential_dict = credential.credential_json
+    access_token = credential_dict["confluence_access_token"]
+
+    try:
+        # Exchange the authorization code for an access token
+        response = requests.get(
+            ConfluenceCloudOAuth.ACCESSIBLE_RESOURCE_URL,
+            headers={
+                "Authorization": f"Bearer {access_token}",
+                "Accept": "application/json",
+            },
+        )
+
+        response.raise_for_status()
+        accessible_resources_data = response.json()
+
+        # Validate the list of AccessibleResources
+        try:
+            accessible_resources = [
+                ConfluenceCloudOAuth.AccessibleResources(**resource)
+                for resource in accessible_resources_data
+            ]
+        except ValidationError as e:
+            raise RuntimeError(f"Failed to parse accessible resources: {e}")
+    except Exception as e:
+        return JSONResponse(
+            status_code=500,
+            content={
+                "success": False,
+                "message": f"An error occurred retrieving Confluence Cloud accessible resources: {str(e)}",
+            },
+        )
+
+    # return the result
+    return JSONResponse(
+        content={
+            "success": True,
+            "message": "Confluence Cloud get accessible resources completed successfully.",
+            "accessible_resources": [
+                resource.model_dump() for resource in accessible_resources
+            ],
+        }
+    )
+
+
+@router.post("/connector/confluence/finalize")
+def confluence_oauth_finalize(
+    credential_id: int,
+    cloud_id: str,
+    cloud_name: str,
+    cloud_url: str,
+    user: User = Depends(current_admin_user),
+    db_session: Session = Depends(get_session),
+    tenant_id: str | None = Depends(get_current_tenant_id),
+) -> JSONResponse:
+    """Saves the info for the selected cloud site to the credential.
+    This is the final step in the confluence oauth flow where after the traditional
+    OAuth process, the user has to select a site to associate with the credentials.
+    After this, the credential is usable."""
+
+    credential = fetch_credential_by_id_for_user(credential_id, user, db_session)
+    if not credential:
+        raise HTTPException(
+            status_code=400,
+            detail=f"Confluence Cloud OAuth failed - credential {credential_id} not found.",
+        )
+
+    new_credential_json: dict[str, Any] = dict(credential.credential_json)
+    new_credential_json["cloud_id"] = cloud_id
+    new_credential_json["cloud_name"] = cloud_name
+    new_credential_json["wiki_base"] = cloud_url
+
+    try:
+        update_credential_json(credential_id, new_credential_json, user, db_session)
+    except Exception as e:
+        return JSONResponse(
+            status_code=500,
+            content={
+                "success": False,
+                "message": f"An error occurred during Confluence Cloud OAuth: {str(e)}",
+            },
+        )
+
+    # return the result
+    return JSONResponse(
+        content={
+            "success": True,
+            "message": "Confluence Cloud OAuth finalized successfully.",
+            "redirect_url": f"{WEB_DOMAIN}/admin/connectors/confluence",
+        }
+    )

backend/ee/onyx/server/oauth/google_drive.py

@@ -0,0 +1,229 @@
+import base64
+import json
+import uuid
+from typing import Any
+from typing import cast
+
+import requests
+from fastapi import Depends
+from fastapi import HTTPException
+from fastapi.responses import JSONResponse
+from pydantic import BaseModel
+from sqlalchemy.orm import Session
+
+from ee.onyx.configs.app_configs import OAUTH_GOOGLE_DRIVE_CLIENT_ID
+from ee.onyx.configs.app_configs import OAUTH_GOOGLE_DRIVE_CLIENT_SECRET
+from ee.onyx.server.oauth.api_router import router
+from onyx.auth.users import current_admin_user
+from onyx.configs.app_configs import DEV_MODE
+from onyx.configs.app_configs import WEB_DOMAIN
+from onyx.configs.constants import DocumentSource
+from onyx.connectors.google_utils.google_auth import get_google_oauth_creds
+from onyx.connectors.google_utils.google_auth import sanitize_oauth_credentials
+from onyx.connectors.google_utils.shared_constants import (
+    DB_CREDENTIALS_AUTHENTICATION_METHOD,
+)
+from onyx.connectors.google_utils.shared_constants import (
+    DB_CREDENTIALS_DICT_TOKEN_KEY,
+)
+from onyx.connectors.google_utils.shared_constants import (
+    DB_CREDENTIALS_PRIMARY_ADMIN_KEY,
+)
+from onyx.connectors.google_utils.shared_constants import (
+    GoogleOAuthAuthenticationMethod,
+)
+from onyx.db.credentials import create_credential
+from onyx.db.engine import get_current_tenant_id
+from onyx.db.engine import get_session
+from onyx.db.models import User
+from onyx.redis.redis_pool import get_redis_client
+from onyx.server.documents.models import CredentialBase
+
+
+class GoogleDriveOAuth:
+    # https://developers.google.com/identity/protocols/oauth2
+    # https://developers.google.com/identity/protocols/oauth2/web-server
+
+    class OAuthSession(BaseModel):
+        """Stored in redis to be looked up on callback"""
+
+        email: str
+        redirect_on_success: str | None  # Where to send the user if OAuth flow succeeds
+
+    CLIENT_ID = OAUTH_GOOGLE_DRIVE_CLIENT_ID
+    CLIENT_SECRET = OAUTH_GOOGLE_DRIVE_CLIENT_SECRET
+
+    TOKEN_URL = "https://oauth2.googleapis.com/token"
+
+    # SCOPE is per https://docs.danswer.dev/connectors/google-drive
+    # TODO: Merge with or use google_utils.GOOGLE_SCOPES
+    SCOPE = (
+        "https://www.googleapis.com/auth/drive.readonly%20"
+        "https://www.googleapis.com/auth/drive.metadata.readonly%20"
+        "https://www.googleapis.com/auth/admin.directory.user.readonly%20"
+        "https://www.googleapis.com/auth/admin.directory.group.readonly"
+    )
+
+    REDIRECT_URI = f"{WEB_DOMAIN}/admin/connectors/google-drive/oauth/callback"
+    DEV_REDIRECT_URI = f"https://redirectmeto.com/{REDIRECT_URI}"
+
+    @classmethod
+    def generate_oauth_url(cls, state: str) -> str:
+        return cls._generate_oauth_url_helper(cls.REDIRECT_URI, state)
+
+    @classmethod
+    def generate_dev_oauth_url(cls, state: str) -> str:
+        """dev mode workaround for localhost testing
+        - https://www.nango.dev/blog/oauth-redirects-on-localhost-with-https
+        """
+
+        return cls._generate_oauth_url_helper(cls.DEV_REDIRECT_URI, state)
+
+    @classmethod
+    def _generate_oauth_url_helper(cls, redirect_uri: str, state: str) -> str:
+        # without prompt=consent, a refresh token is only issued the first time the user approves
+        url = (
+            f"https://accounts.google.com/o/oauth2/v2/auth"
+            f"?client_id={cls.CLIENT_ID}"
+            f"&redirect_uri={redirect_uri}"
+            "&response_type=code"
+            f"&scope={cls.SCOPE}"
+            "&access_type=offline"
+            f"&state={state}"
+            "&prompt=consent"
+        )
+        return url
+
+    @classmethod
+    def session_dump_json(cls, email: str, redirect_on_success: str | None) -> str:
+        """Temporary state to store in redis. to be looked up on auth response.
+        Returns a json string.
+        """
+        session = GoogleDriveOAuth.OAuthSession(
+            email=email, redirect_on_success=redirect_on_success
+        )
+        return session.model_dump_json()
+
+    @classmethod
+    def parse_session(cls, session_json: str) -> OAuthSession:
+        session = GoogleDriveOAuth.OAuthSession.model_validate_json(session_json)
+        return session
+
+
+@router.post("/connector/google-drive/callback")
+def handle_google_drive_oauth_callback(
+    code: str,
+    state: str,
+    user: User = Depends(current_admin_user),
+    db_session: Session = Depends(get_session),
+    tenant_id: str | None = Depends(get_current_tenant_id),
+) -> JSONResponse:
+    if not GoogleDriveOAuth.CLIENT_ID or not GoogleDriveOAuth.CLIENT_SECRET:
+        raise HTTPException(
+            status_code=500,
+            detail="Google Drive client ID or client secret is not configured.",
+        )
+
+    r = get_redis_client(tenant_id=tenant_id)
+
+    # recover the state
+    padded_state = state + "=" * (
+        -len(state) % 4
+    )  # Add padding back (Base64 decoding requires padding)
+    uuid_bytes = base64.urlsafe_b64decode(
+        padded_state
+    )  # Decode the Base64 string back to bytes
+
+    # Convert bytes back to a UUID
+    oauth_uuid = uuid.UUID(bytes=uuid_bytes)
+    oauth_uuid_str = str(oauth_uuid)
+
+    r_key = f"da_oauth:{oauth_uuid_str}"
+
+    session_json_bytes = cast(bytes, r.get(r_key))
+    if not session_json_bytes:
+        raise HTTPException(
+            status_code=400,
+            detail=f"Google Drive OAuth failed - OAuth state key not found: key={r_key}",
+        )
+
+    session_json = session_json_bytes.decode("utf-8")
+    try:
+        session = GoogleDriveOAuth.parse_session(session_json)
+
+        if not DEV_MODE:
+            redirect_uri = GoogleDriveOAuth.REDIRECT_URI
+        else:
+            redirect_uri = GoogleDriveOAuth.DEV_REDIRECT_URI
+
+        # Exchange the authorization code for an access token
+        response = requests.post(
+            GoogleDriveOAuth.TOKEN_URL,
+            headers={"Content-Type": "application/x-www-form-urlencoded"},
+            data={
+                "client_id": GoogleDriveOAuth.CLIENT_ID,
+                "client_secret": GoogleDriveOAuth.CLIENT_SECRET,
+                "code": code,
+                "redirect_uri": redirect_uri,
+                "grant_type": "authorization_code",
+            },
+        )
+
+        response.raise_for_status()
+
+        authorization_response: dict[str, Any] = response.json()
+
+        # the connector wants us to store the json in its authorized_user_info format
+        # returned from OAuthCredentials.get_authorized_user_info().
+        # So refresh immediately via get_google_oauth_creds with the params filled in
+        # from fields in authorization_response to get the json we need
+        authorized_user_info = {}
+        authorized_user_info["client_id"] = OAUTH_GOOGLE_DRIVE_CLIENT_ID
+        authorized_user_info["client_secret"] = OAUTH_GOOGLE_DRIVE_CLIENT_SECRET
+        authorized_user_info["refresh_token"] = authorization_response["refresh_token"]
+
+        token_json_str = json.dumps(authorized_user_info)
+        oauth_creds = get_google_oauth_creds(
+            token_json_str=token_json_str, source=DocumentSource.GOOGLE_DRIVE
+        )
+        if not oauth_creds:
+            raise RuntimeError("get_google_oauth_creds returned None.")
+
+        # save off the credentials
+        oauth_creds_sanitized_json_str = sanitize_oauth_credentials(oauth_creds)
+
+        credential_dict: dict[str, str] = {}
+        credential_dict[DB_CREDENTIALS_DICT_TOKEN_KEY] = oauth_creds_sanitized_json_str
+        credential_dict[DB_CREDENTIALS_PRIMARY_ADMIN_KEY] = session.email
+        credential_dict[
+            DB_CREDENTIALS_AUTHENTICATION_METHOD
+        ] = GoogleOAuthAuthenticationMethod.OAUTH_INTERACTIVE.value
+
+        credential_info = CredentialBase(
+            credential_json=credential_dict,
+            admin_public=True,
+            source=DocumentSource.GOOGLE_DRIVE,
+            name="OAuth (interactive)",
+        )
+
+        create_credential(credential_info, user, db_session)
+    except Exception as e:
+        return JSONResponse(
+            status_code=500,
+            content={
+                "success": False,
+                "message": f"An error occurred during Google Drive OAuth: {str(e)}",
+            },
+        )
+    finally:
+        r.delete(r_key)
+
+    # return the result
+    return JSONResponse(
+        content={
+            "success": True,
+            "message": "Google Drive OAuth completed successfully.",
+            "finalize_url": None,
+            "redirect_on_success": session.redirect_on_success,
+        }
+    )

backend/ee/onyx/server/oauth/slack.py

@@ -0,0 +1,197 @@
+import base64
+import uuid
+from typing import cast
+
+import requests
+from fastapi import Depends
+from fastapi import HTTPException
+from fastapi.responses import JSONResponse
+from pydantic import BaseModel
+from sqlalchemy.orm import Session
+
+from ee.onyx.configs.app_configs import OAUTH_SLACK_CLIENT_ID
+from ee.onyx.configs.app_configs import OAUTH_SLACK_CLIENT_SECRET
+from ee.onyx.server.oauth.api_router import router
+from onyx.auth.users import current_admin_user
+from onyx.configs.app_configs import DEV_MODE
+from onyx.configs.app_configs import WEB_DOMAIN
+from onyx.configs.constants import DocumentSource
+from onyx.db.credentials import create_credential
+from onyx.db.engine import get_current_tenant_id
+from onyx.db.engine import get_session
+from onyx.db.models import User
+from onyx.redis.redis_pool import get_redis_client
+from onyx.server.documents.models import CredentialBase
+
+
+class SlackOAuth:
+    # https://knock.app/blog/how-to-authenticate-users-in-slack-using-oauth
+    # Example: https://api.slack.com/authentication/oauth-v2#exchanging
+
+    class OAuthSession(BaseModel):
+        """Stored in redis to be looked up on callback"""
+
+        email: str
+        redirect_on_success: str | None  # Where to send the user if OAuth flow succeeds
+
+    CLIENT_ID = OAUTH_SLACK_CLIENT_ID
+    CLIENT_SECRET = OAUTH_SLACK_CLIENT_SECRET
+
+    TOKEN_URL = "https://slack.com/api/oauth.v2.access"
+
+    # SCOPE is per https://docs.danswer.dev/connectors/slack
+    BOT_SCOPE = (
+        "channels:history,"
+        "channels:read,"
+        "groups:history,"
+        "groups:read,"
+        "channels:join,"
+        "im:history,"
+        "users:read,"
+        "users:read.email,"
+        "usergroups:read"
+    )
+
+    REDIRECT_URI = f"{WEB_DOMAIN}/admin/connectors/slack/oauth/callback"
+    DEV_REDIRECT_URI = f"https://redirectmeto.com/{REDIRECT_URI}"
+
+    @classmethod
+    def generate_oauth_url(cls, state: str) -> str:
+        return cls._generate_oauth_url_helper(cls.REDIRECT_URI, state)
+
+    @classmethod
+    def generate_dev_oauth_url(cls, state: str) -> str:
+        """dev mode workaround for localhost testing
+        - https://www.nango.dev/blog/oauth-redirects-on-localhost-with-https
+        """
+
+        return cls._generate_oauth_url_helper(cls.DEV_REDIRECT_URI, state)
+
+    @classmethod
+    def _generate_oauth_url_helper(cls, redirect_uri: str, state: str) -> str:
+        url = (
+            f"https://slack.com/oauth/v2/authorize"
+            f"?client_id={cls.CLIENT_ID}"
+            f"&redirect_uri={redirect_uri}"
+            f"&scope={cls.BOT_SCOPE}"
+            f"&state={state}"
+        )
+        return url
+
+    @classmethod
+    def session_dump_json(cls, email: str, redirect_on_success: str | None) -> str:
+        """Temporary state to store in redis. to be looked up on auth response.
+        Returns a json string.
+        """
+        session = SlackOAuth.OAuthSession(
+            email=email, redirect_on_success=redirect_on_success
+        )
+        return session.model_dump_json()
+
+    @classmethod
+    def parse_session(cls, session_json: str) -> OAuthSession:
+        session = SlackOAuth.OAuthSession.model_validate_json(session_json)
+        return session
+
+
+@router.post("/connector/slack/callback")
+def handle_slack_oauth_callback(
+    code: str,
+    state: str,
+    user: User = Depends(current_admin_user),
+    db_session: Session = Depends(get_session),
+    tenant_id: str | None = Depends(get_current_tenant_id),
+) -> JSONResponse:
+    if not SlackOAuth.CLIENT_ID or not SlackOAuth.CLIENT_SECRET:
+        raise HTTPException(
+            status_code=500,
+            detail="Slack client ID or client secret is not configured.",
+        )
+
+    r = get_redis_client(tenant_id=tenant_id)
+
+    # recover the state
+    padded_state = state + "=" * (
+        -len(state) % 4
+    )  # Add padding back (Base64 decoding requires padding)
+    uuid_bytes = base64.urlsafe_b64decode(
+        padded_state
+    )  # Decode the Base64 string back to bytes
+
+    # Convert bytes back to a UUID
+    oauth_uuid = uuid.UUID(bytes=uuid_bytes)
+    oauth_uuid_str = str(oauth_uuid)
+
+    r_key = f"da_oauth:{oauth_uuid_str}"
+
+    session_json_bytes = cast(bytes, r.get(r_key))
+    if not session_json_bytes:
+        raise HTTPException(
+            status_code=400,
+            detail=f"Slack OAuth failed - OAuth state key not found: key={r_key}",
+        )
+
+    session_json = session_json_bytes.decode("utf-8")
+    try:
+        session = SlackOAuth.parse_session(session_json)
+
+        if not DEV_MODE:
+            redirect_uri = SlackOAuth.REDIRECT_URI
+        else:
+            redirect_uri = SlackOAuth.DEV_REDIRECT_URI
+
+        # Exchange the authorization code for an access token
+        response = requests.post(
+            SlackOAuth.TOKEN_URL,
+            headers={"Content-Type": "application/x-www-form-urlencoded"},
+            data={
+                "client_id": SlackOAuth.CLIENT_ID,
+                "client_secret": SlackOAuth.CLIENT_SECRET,
+                "code": code,
+                "redirect_uri": redirect_uri,
+            },
+        )
+
+        response_data = response.json()
+
+        if not response_data.get("ok"):
+            raise HTTPException(
+                status_code=400,
+                detail=f"Slack OAuth failed: {response_data.get('error')}",
+            )
+
+        # Extract token and team information
+        access_token: str = response_data.get("access_token")
+        team_id: str = response_data.get("team", {}).get("id")
+        authed_user_id: str = response_data.get("authed_user", {}).get("id")
+
+        credential_info = CredentialBase(
+            credential_json={"slack_bot_token": access_token},
+            admin_public=True,
+            source=DocumentSource.SLACK,
+            name="Slack OAuth",
+        )
+
+        create_credential(credential_info, user, db_session)
+    except Exception as e:
+        return JSONResponse(
+            status_code=500,
+            content={
+                "success": False,
+                "message": f"An error occurred during Slack OAuth: {str(e)}",
+            },
+        )
+    finally:
+        r.delete(r_key)
+
+    # return the result
+    return JSONResponse(
+        content={
+            "success": True,
+            "message": "Slack OAuth completed successfully.",
+            "finalize_url": None,
+            "redirect_on_success": session.redirect_on_success,
+            "team_id": team_id,
+            "authed_user_id": authed_user_id,
+        }
+    )

backend/ee/onyx/server/query_and_chat/token_limit.py

@@ -13,7 +13,7 @@ from sqlalchemy import select
 from sqlalchemy.orm import Session
 
 from onyx.db.api_key import is_api_key_email_address
-from onyx.db.engine import get_session_with_tenant
+from onyx.db.engine import get_session_with_current_tenant
 from onyx.db.models import ChatMessage
 from onyx.db.models import ChatSession
 from onyx.db.models import TokenRateLimit
@@ -28,21 +28,21 @@ from onyx.server.query_and_chat.token_limit import _user_is_rate_limited_by_glob
 from onyx.utils.threadpool_concurrency import run_functions_tuples_in_parallel
 
 
-def _check_token_rate_limits(user: User | None, tenant_id: str | None) -> None:
+def _check_token_rate_limits(user: User | None) -> None:
     if user is None:
         # Unauthenticated users are only rate limited by global settings
-        _user_is_rate_limited_by_global(tenant_id)
+        _user_is_rate_limited_by_global()
 
     elif is_api_key_email_address(user.email):
         # API keys are only rate limited by global settings
-        _user_is_rate_limited_by_global(tenant_id)
+        _user_is_rate_limited_by_global()
 
     else:
         run_functions_tuples_in_parallel(
             [
-                (_user_is_rate_limited, (user.id, tenant_id)),
-                (_user_is_rate_limited_by_group, (user.id, tenant_id)),
-                (_user_is_rate_limited_by_global, (tenant_id,)),
+                (_user_is_rate_limited, (user.id,)),
+                (_user_is_rate_limited_by_group, (user.id,)),
+                (_user_is_rate_limited_by_global, ()),
             ]
         )
 
@@ -52,8 +52,8 @@ User rate limits
 """
 
 
-def _user_is_rate_limited(user_id: UUID, tenant_id: str | None) -> None:
-    with get_session_with_tenant(tenant_id) as db_session:
+def _user_is_rate_limited(user_id: UUID) -> None:
+    with get_session_with_current_tenant() as db_session:
         user_rate_limits = fetch_all_user_token_rate_limits(
             db_session=db_session, enabled_only=True, ordered=False
         )
@@ -93,8 +93,8 @@ User Group rate limits
 """
 
 
-def _user_is_rate_limited_by_group(user_id: UUID, tenant_id: str | None) -> None:
-    with get_session_with_tenant(tenant_id) as db_session:
+def _user_is_rate_limited_by_group(user_id: UUID) -> None:
+    with get_session_with_current_tenant() as db_session:
         group_rate_limits = _fetch_all_user_group_rate_limits(user_id, db_session)
 
         if group_rate_limits:

backend/ee/onyx/server/query_history/api.py

@@ -2,6 +2,7 @@ import csv
 import io
 from datetime import datetime
 from datetime import timezone
+from http import HTTPStatus
 from uuid import UUID
 
 from fastapi import APIRouter
@@ -21,8 +22,10 @@ from ee.onyx.server.query_history.models import QuestionAnswerPairSnapshot
 from onyx.auth.users import current_admin_user
 from onyx.auth.users import get_display_email
 from onyx.chat.chat_utils import create_chat_chain
+from onyx.configs.app_configs import ONYX_QUERY_HISTORY_TYPE
 from onyx.configs.constants import MessageType
 from onyx.configs.constants import QAFeedbackType
+from onyx.configs.constants import QueryHistoryType
 from onyx.configs.constants import SessionType
 from onyx.db.chat import get_chat_session_by_id
 from onyx.db.chat import get_chat_sessions_by_user
@@ -35,6 +38,8 @@ from onyx.server.query_and_chat.models import ChatSessionsResponse
 
 router = APIRouter()
 
+ONYX_ANONYMIZED_EMAIL = "anonymous@anonymous.invalid"
+
 
 def fetch_and_process_chat_session_history(
     db_session: Session,
@@ -107,6 +112,17 @@ def get_user_chat_sessions(
     _: User | None = Depends(current_admin_user),
     db_session: Session = Depends(get_session),
 ) -> ChatSessionsResponse:
+    # we specifically don't allow this endpoint if "anonymized" since
+    # this is a direct query on the user id
+    if ONYX_QUERY_HISTORY_TYPE in [
+        QueryHistoryType.DISABLED,
+        QueryHistoryType.ANONYMIZED,
+    ]:
+        raise HTTPException(
+            status_code=HTTPStatus.FORBIDDEN,
+            detail="Per user query history has been disabled by the administrator.",
+        )
+
     try:
         chat_sessions = get_chat_sessions_by_user(
             user_id=user_id, deleted=False, db_session=db_session, limit=0
@@ -122,6 +138,7 @@ def get_user_chat_sessions(
                 name=chat.description,
                 persona_id=chat.persona_id,
                 time_created=chat.time_created.isoformat(),
+                time_updated=chat.time_updated.isoformat(),
                 shared_status=chat.shared_status,
                 folder_id=chat.folder_id,
                 current_alternate_model=chat.current_alternate_model,
@@ -141,6 +158,12 @@ def get_chat_session_history(
     _: User | None = Depends(current_admin_user),
     db_session: Session = Depends(get_session),
 ) -> PaginatedReturn[ChatSessionMinimal]:
+    if ONYX_QUERY_HISTORY_TYPE == QueryHistoryType.DISABLED:
+        raise HTTPException(
+            status_code=HTTPStatus.FORBIDDEN,
+            detail="Query history has been disabled by the administrator.",
+        )
+
     page_of_chat_sessions = get_page_of_chat_sessions(
         page_num=page_num,
         page_size=page_size,
@@ -157,11 +180,16 @@ def get_chat_session_history(
         feedback_filter=feedback_type,
     )
 
+    minimal_chat_sessions: list[ChatSessionMinimal] = []
+
+    for chat_session in page_of_chat_sessions:
+        minimal_chat_session = ChatSessionMinimal.from_chat_session(chat_session)
+        if ONYX_QUERY_HISTORY_TYPE == QueryHistoryType.ANONYMIZED:
+            minimal_chat_session.user_email = ONYX_ANONYMIZED_EMAIL
+        minimal_chat_sessions.append(minimal_chat_session)
+
     return PaginatedReturn(
-        items=[
-            ChatSessionMinimal.from_chat_session(chat_session)
-            for chat_session in page_of_chat_sessions
-        ],
+        items=minimal_chat_sessions,
         total_items=total_filtered_chat_sessions_count,
     )
 
@@ -172,6 +200,12 @@ def get_chat_session_admin(
     _: User | None = Depends(current_admin_user),
     db_session: Session = Depends(get_session),
 ) -> ChatSessionSnapshot:
+    if ONYX_QUERY_HISTORY_TYPE == QueryHistoryType.DISABLED:
+        raise HTTPException(
+            status_code=HTTPStatus.FORBIDDEN,
+            detail="Query history has been disabled by the administrator.",
+        )
+
     try:
         chat_session = get_chat_session_by_id(
             chat_session_id=chat_session_id,
@@ -193,6 +227,9 @@ def get_chat_session_admin(
             f"Could not create snapshot for chat session with id '{chat_session_id}'",
         )
 
+    if ONYX_QUERY_HISTORY_TYPE == QueryHistoryType.ANONYMIZED:
+        snapshot.user_email = ONYX_ANONYMIZED_EMAIL
+
     return snapshot
 
 
@@ -203,6 +240,12 @@ def get_query_history_as_csv(
     end: datetime | None = None,
     db_session: Session = Depends(get_session),
 ) -> StreamingResponse:
+    if ONYX_QUERY_HISTORY_TYPE == QueryHistoryType.DISABLED:
+        raise HTTPException(
+            status_code=HTTPStatus.FORBIDDEN,
+            detail="Query history has been disabled by the administrator.",
+        )
+
     complete_chat_session_history = fetch_and_process_chat_session_history(
         db_session=db_session,
         start=start or datetime.fromtimestamp(0, tz=timezone.utc),
@@ -213,6 +256,9 @@ def get_query_history_as_csv(
 
     question_answer_pairs: list[QuestionAnswerPairSnapshot] = []
     for chat_session_snapshot in complete_chat_session_history:
+        if ONYX_QUERY_HISTORY_TYPE == QueryHistoryType.ANONYMIZED:
+            chat_session_snapshot.user_email = ONYX_ANONYMIZED_EMAIL
+
         question_answer_pairs.extend(
             QuestionAnswerPairSnapshot.from_chat_session_snapshot(chat_session_snapshot)
         )

backend/ee/onyx/server/tenants/api.py

@@ -41,14 +41,15 @@ from onyx.auth.users import User
 from onyx.configs.app_configs import WEB_DOMAIN
 from onyx.configs.constants import FASTAPI_USERS_AUTH_COOKIE_NAME
 from onyx.db.auth import get_user_count
-from onyx.db.engine import get_current_tenant_id
 from onyx.db.engine import get_session
+from onyx.db.engine import get_session_with_shared_schema
 from onyx.db.engine import get_session_with_tenant
 from onyx.db.users import delete_user_from_db
 from onyx.db.users import get_user_by_email
 from onyx.server.manage.models import UserByEmail
 from onyx.utils.logger import setup_logger
 from shared_configs.contextvars import CURRENT_TENANT_ID_CONTEXTVAR
+from shared_configs.contextvars import get_current_tenant_id
 
 stripe.api_key = STRIPE_SECRET_KEY
 logger = setup_logger()
@@ -57,13 +58,14 @@ router = APIRouter(prefix="/tenants")
 
 @router.get("/anonymous-user-path")
 async def get_anonymous_user_path_api(
-    tenant_id: str | None = Depends(get_current_tenant_id),
     _: User | None = Depends(current_admin_user),
 ) -> AnonymousUserPath:
+    tenant_id = get_current_tenant_id()
+
     if tenant_id is None:
         raise HTTPException(status_code=404, detail="Tenant not found")
 
-    with get_session_with_tenant(tenant_id=None) as db_session:
+    with get_session_with_shared_schema() as db_session:
         current_path = get_anonymous_user_path(tenant_id, db_session)
 
     return AnonymousUserPath(anonymous_user_path=current_path)
@@ -72,15 +74,15 @@ async def get_anonymous_user_path_api(
 @router.post("/anonymous-user-path")
 async def set_anonymous_user_path_api(
     anonymous_user_path: str,
-    tenant_id: str = Depends(get_current_tenant_id),
     _: User | None = Depends(current_admin_user),
 ) -> None:
+    tenant_id = get_current_tenant_id()
     try:
         validate_anonymous_user_path(anonymous_user_path)
     except ValueError as e:
         raise HTTPException(status_code=400, detail=str(e))
 
-    with get_session_with_tenant(tenant_id=None) as db_session:
+    with get_session_with_shared_schema() as db_session:
         try:
             modify_anonymous_user_path(tenant_id, anonymous_user_path, db_session)
         except IntegrityError:
@@ -101,7 +103,7 @@ async def login_as_anonymous_user(
     anonymous_user_path: str,
     _: User | None = Depends(optional_user),
 ) -> Response:
-    with get_session_with_tenant(tenant_id=None) as db_session:
+    with get_session_with_shared_schema() as db_session:
         tenant_id = get_tenant_id_for_anonymous_user_path(
             anonymous_user_path, db_session
         )
@@ -150,14 +152,17 @@ async def billing_information(
     _: User = Depends(current_admin_user),
 ) -> BillingInformation | SubscriptionStatusResponse:
     logger.info("Fetching billing information")
-    return fetch_billing_information(CURRENT_TENANT_ID_CONTEXTVAR.get())
+    tenant_id = get_current_tenant_id()
+    return fetch_billing_information(tenant_id)
 
 
 @router.post("/create-customer-portal-session")
-async def create_customer_portal_session(_: User = Depends(current_admin_user)) -> dict:
+async def create_customer_portal_session(
+    _: User = Depends(current_admin_user),
+) -> dict:
+    tenant_id = get_current_tenant_id()
+
     try:
-        # Fetch tenant_id and current tenant's information
-        tenant_id = CURRENT_TENANT_ID_CONTEXTVAR.get()
         stripe_info = fetch_tenant_stripe_information(tenant_id)
         stripe_customer_id = stripe_info.get("stripe_customer_id")
         if not stripe_customer_id:
@@ -181,6 +186,8 @@ async def create_subscription_session(
 ) -> SubscriptionSessionResponse:
     try:
         tenant_id = CURRENT_TENANT_ID_CONTEXTVAR.get()
+        if not tenant_id:
+            raise HTTPException(status_code=400, detail="Tenant ID not found")
         session_id = fetch_stripe_checkout_session(tenant_id)
         return SubscriptionSessionResponse(sessionId=session_id)
 
@@ -197,7 +204,7 @@ async def impersonate_user(
     """Allows a cloud superuser to impersonate another user by generating an impersonation JWT token"""
     tenant_id = get_tenant_id_for_email(impersonate_request.email)
 
-    with get_session_with_tenant(tenant_id) as tenant_session:
+    with get_session_with_tenant(tenant_id=tenant_id) as tenant_session:
         user_to_impersonate = get_user_by_email(
             impersonate_request.email, tenant_session
         )
@@ -221,8 +228,9 @@ async def leave_organization(
     user_email: UserByEmail,
     current_user: User | None = Depends(current_admin_user),
     db_session: Session = Depends(get_session),
-    tenant_id: str = Depends(get_current_tenant_id),
 ) -> None:
+    tenant_id = get_current_tenant_id()
+
     if current_user is None or current_user.email != user_email.user_email:
         raise HTTPException(
             status_code=403, detail="You can only leave the organization as yourself"

backend/ee/onyx/server/tenants/billing.py

@@ -7,6 +7,7 @@ from ee.onyx.configs.app_configs import STRIPE_PRICE_ID
 from ee.onyx.configs.app_configs import STRIPE_SECRET_KEY
 from ee.onyx.server.tenants.access import generate_data_plane_token
 from ee.onyx.server.tenants.models import BillingInformation
+from ee.onyx.server.tenants.models import SubscriptionStatusResponse
 from onyx.configs.app_configs import CONTROL_PLANE_API_BASE_URL
 from onyx.utils.logger import setup_logger
 
@@ -41,7 +42,9 @@ def fetch_tenant_stripe_information(tenant_id: str) -> dict:
     return response.json()
 
 
-def fetch_billing_information(tenant_id: str) -> BillingInformation:
+def fetch_billing_information(
+    tenant_id: str,
+) -> BillingInformation | SubscriptionStatusResponse:
     logger.info("Fetching billing information")
     token = generate_data_plane_token()
     headers = {
@@ -52,8 +55,19 @@ def fetch_billing_information(tenant_id: str) -> BillingInformation:
     params = {"tenant_id": tenant_id}
     response = requests.get(url, headers=headers, params=params)
     response.raise_for_status()
-    billing_info = BillingInformation(**response.json())
-    return billing_info
+
+    response_data = response.json()
+
+    # Check if the response indicates no subscription
+    if (
+        isinstance(response_data, dict)
+        and "subscribed" in response_data
+        and not response_data["subscribed"]
+    ):
+        return SubscriptionStatusResponse(**response_data)
+
+    # Otherwise, parse as BillingInformation
+    return BillingInformation(**response_data)
 
 
 def register_tenant_users(tenant_id: str, number_of_users: int) -> stripe.Subscription:

backend/ee/onyx/server/tenants/product_gating.py

@@ -48,4 +48,5 @@ def store_product_gating(tenant_id: str, application_status: ApplicationStatus)
 
 def get_gated_tenants() -> set[str]:
     redis_client = get_redis_replica_client(tenant_id=ONYX_CLOUD_TENANT_ID)
-    return cast(set[str], redis_client.smembers(GATED_TENANTS_KEY))
+    gated_tenants_bytes = cast(set[bytes], redis_client.smembers(GATED_TENANTS_KEY))
+    return {tenant_id.decode("utf-8") for tenant_id in gated_tenants_bytes}

backend/ee/onyx/server/tenants/provisioning.py

@@ -55,7 +55,11 @@ logger = logging.getLogger(__name__)
 async def get_or_provision_tenant(
     email: str, referral_source: str | None = None, request: Request | None = None
 ) -> str:
-    """Get existing tenant ID for an email or create a new tenant if none exists."""
+    """
+    Get existing tenant ID for an email or create a new tenant if none exists.
+    This function should only be called after we have verified we want this user's tenant to exist.
+    It returns the tenant ID associated with the email, creating a new tenant if necessary.
+    """
     if not MULTI_TENANT:
         return POSTGRES_DEFAULT_SCHEMA
 
@@ -104,21 +108,21 @@ async def provision_tenant(tenant_id: str, email: str) -> None:
             status_code=409, detail="User already belongs to an organization"
         )
 
-    logger.info(f"Provisioning tenant: {tenant_id}")
+    logger.debug(f"Provisioning tenant {tenant_id} for user {email}")
     token = None
 
     try:
         if not create_schema_if_not_exists(tenant_id):
-            logger.info(f"Created schema for tenant {tenant_id}")
+            logger.debug(f"Created schema for tenant {tenant_id}")
         else:
-            logger.info(f"Schema already exists for tenant {tenant_id}")
+            logger.debug(f"Schema already exists for tenant {tenant_id}")
 
         token = CURRENT_TENANT_ID_CONTEXTVAR.set(tenant_id)
 
         # Await the Alembic migrations
         await asyncio.to_thread(run_alembic_migrations, tenant_id)
 
-        with get_session_with_tenant(tenant_id) as db_session:
+        with get_session_with_tenant(tenant_id=tenant_id) as db_session:
             configure_default_api_keys(db_session)
 
             current_search_settings = (
@@ -134,7 +138,7 @@ async def provision_tenant(tenant_id: str, email: str) -> None:
 
         add_users_to_tenant([email], tenant_id)
 
-        with get_session_with_tenant(tenant_id) as db_session:
+        with get_session_with_tenant(tenant_id=tenant_id) as db_session:
             create_milestone_and_report(
                 user=None,
                 distinct_id=tenant_id,
@@ -200,33 +204,15 @@ async def rollback_tenant_provisioning(tenant_id: str) -> None:
 
 
 def configure_default_api_keys(db_session: Session) -> None:
-    if OPENAI_DEFAULT_API_KEY:
-        open_provider = LLMProviderUpsertRequest(
-            name="OpenAI",
-            provider=OPENAI_PROVIDER_NAME,
-            api_key=OPENAI_DEFAULT_API_KEY,
-            default_model_name="gpt-4",
-            fast_default_model_name="gpt-4o-mini",
-            model_names=OPEN_AI_MODEL_NAMES,
-        )
-        try:
-            full_provider = upsert_llm_provider(open_provider, db_session)
-            update_default_provider(full_provider.id, db_session)
-        except Exception as e:
-            logger.error(f"Failed to configure OpenAI provider: {e}")
-    else:
-        logger.error(
-            "OPENAI_DEFAULT_API_KEY not set, skipping OpenAI provider configuration"
-        )
-
     if ANTHROPIC_DEFAULT_API_KEY:
         anthropic_provider = LLMProviderUpsertRequest(
             name="Anthropic",
             provider=ANTHROPIC_PROVIDER_NAME,
             api_key=ANTHROPIC_DEFAULT_API_KEY,
-            default_model_name="claude-3-5-sonnet-20241022",
+            default_model_name="claude-3-7-sonnet-20250219",
             fast_default_model_name="claude-3-5-sonnet-20241022",
             model_names=ANTHROPIC_MODEL_NAMES,
+            display_model_names=["claude-3-5-sonnet-20241022"],
         )
         try:
             full_provider = upsert_llm_provider(anthropic_provider, db_session)
@@ -238,6 +224,26 @@ def configure_default_api_keys(db_session: Session) -> None:
             "ANTHROPIC_DEFAULT_API_KEY not set, skipping Anthropic provider configuration"
         )
 
+    if OPENAI_DEFAULT_API_KEY:
+        open_provider = LLMProviderUpsertRequest(
+            name="OpenAI",
+            provider=OPENAI_PROVIDER_NAME,
+            api_key=OPENAI_DEFAULT_API_KEY,
+            default_model_name="gpt-4o",
+            fast_default_model_name="gpt-4o-mini",
+            model_names=OPEN_AI_MODEL_NAMES,
+            display_model_names=["o1", "o3-mini", "gpt-4o", "gpt-4o-mini"],
+        )
+        try:
+            full_provider = upsert_llm_provider(open_provider, db_session)
+            update_default_provider(full_provider.id, db_session)
+        except Exception as e:
+            logger.error(f"Failed to configure OpenAI provider: {e}")
+    else:
+        logger.error(
+            "OPENAI_DEFAULT_API_KEY not set, skipping OpenAI provider configuration"
+        )
+
     if COHERE_DEFAULT_API_KEY:
         cloud_embedding_provider = CloudEmbeddingProviderCreationRequest(
             provider_type=EmbeddingProvider.COHERE,

backend/ee/onyx/server/tenants/user_mapping.py

@@ -28,7 +28,7 @@ def get_tenant_id_for_email(email: str) -> str:
 
 
 def user_owns_a_tenant(email: str) -> bool:
-    with get_session_with_tenant(POSTGRES_DEFAULT_SCHEMA) as db_session:
+    with get_session_with_tenant(tenant_id=POSTGRES_DEFAULT_SCHEMA) as db_session:
         result = (
             db_session.query(UserTenantMapping)
             .filter(UserTenantMapping.email == email)
@@ -38,7 +38,7 @@ def user_owns_a_tenant(email: str) -> bool:
 
 
 def add_users_to_tenant(emails: list[str], tenant_id: str) -> None:
-    with get_session_with_tenant(POSTGRES_DEFAULT_SCHEMA) as db_session:
+    with get_session_with_tenant(tenant_id=POSTGRES_DEFAULT_SCHEMA) as db_session:
         try:
             for email in emails:
                 db_session.add(UserTenantMapping(email=email, tenant_id=tenant_id))
@@ -48,7 +48,7 @@ def add_users_to_tenant(emails: list[str], tenant_id: str) -> None:
 
 
 def remove_users_from_tenant(emails: list[str], tenant_id: str) -> None:
-    with get_session_with_tenant(POSTGRES_DEFAULT_SCHEMA) as db_session:
+    with get_session_with_tenant(tenant_id=POSTGRES_DEFAULT_SCHEMA) as db_session:
         try:
             mappings_to_delete = (
                 db_session.query(UserTenantMapping)
@@ -71,7 +71,7 @@ def remove_users_from_tenant(emails: list[str], tenant_id: str) -> None:
 
 
 def remove_all_users_from_tenant(tenant_id: str) -> None:
-    with get_session_with_tenant(POSTGRES_DEFAULT_SCHEMA) as db_session:
+    with get_session_with_tenant(tenant_id=POSTGRES_DEFAULT_SCHEMA) as db_session:
         db_session.query(UserTenantMapping).filter(
             UserTenantMapping.tenant_id == tenant_id
         ).delete()

backend/model_server/constants.py

@@ -6,7 +6,7 @@ MODEL_WARM_UP_STRING = "hi " * 512
 DEFAULT_OPENAI_MODEL = "text-embedding-3-small"
 DEFAULT_COHERE_MODEL = "embed-english-light-v3.0"
 DEFAULT_VOYAGE_MODEL = "voyage-large-2-instruct"
-DEFAULT_VERTEX_MODEL = "text-embedding-004"
+DEFAULT_VERTEX_MODEL = "text-embedding-005"
 
 
 class EmbeddingModelTextType:

backend/model_server/encoders.py

@@ -5,6 +5,7 @@ from types import TracebackType
 from typing import cast
 from typing import Optional
 
+import aioboto3  # type: ignore
 import httpx
 import openai
 import vertexai  # type: ignore
@@ -28,11 +29,13 @@ from model_server.constants import DEFAULT_VERTEX_MODEL
 from model_server.constants import DEFAULT_VOYAGE_MODEL
 from model_server.constants import EmbeddingModelTextType
 from model_server.constants import EmbeddingProvider
+from model_server.utils import pass_aws_key
 from model_server.utils import simple_log_function_time
 from onyx.utils.logger import setup_logger
 from shared_configs.configs import API_BASED_EMBEDDING_TIMEOUT
 from shared_configs.configs import INDEXING_ONLY
 from shared_configs.configs import OPENAI_EMBEDDING_TIMEOUT
+from shared_configs.configs import VERTEXAI_EMBEDDING_LOCAL_BATCH_SIZE
 from shared_configs.enums import EmbedTextType
 from shared_configs.enums import RerankerProvider
 from shared_configs.model_server_models import Embedding
@@ -78,7 +81,7 @@ class CloudEmbedding:
         self._closed = False
 
     async def _embed_openai(
-        self, texts: list[str], model: str | None
+        self, texts: list[str], model: str | None, reduced_dimension: int | None
     ) -> list[Embedding]:
         if not model:
             model = DEFAULT_OPENAI_MODEL
@@ -91,19 +94,28 @@ class CloudEmbedding:
         final_embeddings: list[Embedding] = []
         try:
             for text_batch in batch_list(texts, _OPENAI_MAX_INPUT_LEN):
-                response = await client.embeddings.create(input=text_batch, model=model)
+                response = await client.embeddings.create(
+                    input=text_batch,
+                    model=model,
+                    dimensions=reduced_dimension or openai.NOT_GIVEN,
+                )
                 final_embeddings.extend(
                     [embedding.embedding for embedding in response.data]
                 )
             return final_embeddings
         except Exception as e:
             error_string = (
-                f"Error embedding text with OpenAI: {str(e)} \n"
-                f"Model: {model} \n"
-                f"Provider: {self.provider} \n"
-                f"Texts: {texts}"
+                f"Exception embedding text with OpenAI - {type(e)}: "
+                f"Model: {model} "
+                f"Provider: {self.provider} "
+                f"Exception: {e}"
             )
             logger.error(error_string)
+
+            # only log text when it's not an authentication error.
+            if not isinstance(e, openai.AuthenticationError):
+                logger.debug(f"Exception texts: {texts}")
+
             raise RuntimeError(error_string)
 
     async def _embed_cohere(
@@ -173,17 +185,24 @@ class CloudEmbedding:
         vertexai.init(project=project_id, credentials=credentials)
         client = TextEmbeddingModel.from_pretrained(model)
 
-        embeddings = await client.get_embeddings_async(
-            [
-                TextEmbeddingInput(
-                    text,
-                    embedding_type,
-                )
-                for text in texts
-            ],
-            auto_truncate=True,  # This is the default
-        )
-        return [embedding.values for embedding in embeddings]
+        inputs = [TextEmbeddingInput(text, embedding_type) for text in texts]
+
+        # Split into batches of 25 texts
+        max_texts_per_batch = VERTEXAI_EMBEDDING_LOCAL_BATCH_SIZE
+        batches = [
+            inputs[i : i + max_texts_per_batch]
+            for i in range(0, len(inputs), max_texts_per_batch)
+        ]
+
+        # Dispatch all embedding calls asynchronously at once
+        tasks = [
+            client.get_embeddings_async(batch, auto_truncate=True) for batch in batches
+        ]
+
+        # Wait for all tasks to complete in parallel
+        results = await asyncio.gather(*tasks)
+
+        return [embedding.values for batch in results for embedding in batch]
 
     async def _embed_litellm_proxy(
         self, texts: list[str], model_name: str | None
@@ -218,9 +237,10 @@ class CloudEmbedding:
         text_type: EmbedTextType,
         model_name: str | None = None,
         deployment_name: str | None = None,
+        reduced_dimension: int | None = None,
     ) -> list[Embedding]:
         if self.provider == EmbeddingProvider.OPENAI:
-            return await self._embed_openai(texts, model_name)
+            return await self._embed_openai(texts, model_name, reduced_dimension)
         elif self.provider == EmbeddingProvider.AZURE:
             return await self._embed_azure(texts, f"azure/{deployment_name}")
         elif self.provider == EmbeddingProvider.LITELLM:
@@ -321,6 +341,7 @@ async def embed_text(
     prefix: str | None,
     api_url: str | None,
     api_version: str | None,
+    reduced_dimension: int | None,
     gpu_type: str = "UNKNOWN",
 ) -> list[Embedding]:
     if not all(texts):
@@ -364,6 +385,7 @@ async def embed_text(
                 model_name=model_name,
                 deployment_name=deployment_name,
                 text_type=text_type,
+                reduced_dimension=reduced_dimension,
             )
 
         if any(embedding is None for embedding in embeddings):
@@ -435,7 +457,7 @@ async def local_rerank(query: str, docs: list[str], model_name: str) -> list[flo
     )
 
 
-async def cohere_rerank(
+async def cohere_rerank_api(
     query: str, docs: list[str], model_name: str, api_key: str
 ) -> list[float]:
     cohere_client = CohereAsyncClient(api_key=api_key)
@@ -445,6 +467,45 @@ async def cohere_rerank(
     return [result.relevance_score for result in sorted_results]
 
 
+async def cohere_rerank_aws(
+    query: str,
+    docs: list[str],
+    model_name: str,
+    region_name: str,
+    aws_access_key_id: str,
+    aws_secret_access_key: str,
+) -> list[float]:
+    session = aioboto3.Session(
+        aws_access_key_id=aws_access_key_id, aws_secret_access_key=aws_secret_access_key
+    )
+    async with session.client(
+        "bedrock-runtime", region_name=region_name
+    ) as bedrock_client:
+        body = json.dumps(
+            {
+                "query": query,
+                "documents": docs,
+                "api_version": 2,
+            }
+        )
+        # Invoke the Bedrock model asynchronously
+        response = await bedrock_client.invoke_model(
+            modelId=model_name,
+            accept="application/json",
+            contentType="application/json",
+            body=body,
+        )
+
+        # Read the response asynchronously
+        response_body = json.loads(await response["body"].read())
+
+        # Extract and sort the results
+        results = response_body.get("results", [])
+        sorted_results = sorted(results, key=lambda item: item["index"])
+
+        return [result["relevance_score"] for result in sorted_results]
+
+
 async def litellm_rerank(
     query: str, docs: list[str], api_url: str, model_name: str, api_key: str | None
 ) -> list[float]:
@@ -503,6 +564,7 @@ async def process_embed_request(
             text_type=embed_request.text_type,
             api_url=embed_request.api_url,
             api_version=embed_request.api_version,
+            reduced_dimension=embed_request.reduced_dimension,
             prefix=prefix,
             gpu_type=gpu_type,
         )
@@ -559,15 +621,32 @@ async def process_rerank_request(rerank_request: RerankRequest) -> RerankRespons
         elif rerank_request.provider_type == RerankerProvider.COHERE:
             if rerank_request.api_key is None:
                 raise RuntimeError("Cohere Rerank Requires an API Key")
-            sim_scores = await cohere_rerank(
+            sim_scores = await cohere_rerank_api(
                 query=rerank_request.query,
                 docs=rerank_request.documents,
                 model_name=rerank_request.model_name,
                 api_key=rerank_request.api_key,
             )
             return RerankResponse(scores=sim_scores)
+
+        elif rerank_request.provider_type == RerankerProvider.BEDROCK:
+            if rerank_request.api_key is None:
+                raise RuntimeError("Bedrock Rerank Requires an API Key")
+            aws_access_key_id, aws_secret_access_key, aws_region = pass_aws_key(
+                rerank_request.api_key
+            )
+            sim_scores = await cohere_rerank_aws(
+                query=rerank_request.query,
+                docs=rerank_request.documents,
+                model_name=rerank_request.model_name,
+                region_name=aws_region,
+                aws_access_key_id=aws_access_key_id,
+                aws_secret_access_key=aws_secret_access_key,
+            )
+            return RerankResponse(scores=sim_scores)
         else:
             raise ValueError(f"Unsupported provider: {rerank_request.provider_type}")
+
     except Exception as e:
         logger.exception(f"Error during reranking process:\n{str(e)}")
         raise HTTPException(

backend/model_server/utils.py

@@ -70,3 +70,32 @@ def get_gpu_type() -> str:
         return GPUStatus.MAC_MPS
 
     return GPUStatus.NONE
+
+
+def pass_aws_key(api_key: str) -> tuple[str, str, str]:
+    """Parse AWS API key string into components.
+
+    Args:
+        api_key: String in format 'aws_ACCESSKEY_SECRETKEY_REGION'
+
+    Returns:
+        Tuple of (access_key, secret_key, region)
+
+    Raises:
+        ValueError: If key format is invalid
+    """
+    if not api_key.startswith("aws"):
+        raise ValueError("API key must start with 'aws' prefix")
+
+    parts = api_key.split("_")
+    if len(parts) != 4:
+        raise ValueError(
+            f"API key must be in format 'aws_ACCESSKEY_SECRETKEY_REGION', got {len(parts) - 1} parts"
+            "this is an onyx specific format for formatting the aws secrets for bedrock"
+        )
+
+    try:
+        _, aws_access_key_id, aws_secret_access_key, aws_region = parts
+        return aws_access_key_id, aws_secret_access_key, aws_region
+    except Exception as e:
+        raise ValueError(f"Failed to parse AWS key components: {str(e)}")

backend/onyx/agents/agent_search/basic/graph_builder.py

@@ -5,14 +5,14 @@ from langgraph.graph import StateGraph
 from onyx.agents.agent_search.basic.states import BasicInput
 from onyx.agents.agent_search.basic.states import BasicOutput
 from onyx.agents.agent_search.basic.states import BasicState
-from onyx.agents.agent_search.orchestration.nodes.basic_use_tool_response import (
-    basic_use_tool_response,
-)
-from onyx.agents.agent_search.orchestration.nodes.llm_tool_choice import llm_tool_choice
+from onyx.agents.agent_search.orchestration.nodes.call_tool import call_tool
+from onyx.agents.agent_search.orchestration.nodes.choose_tool import choose_tool
 from onyx.agents.agent_search.orchestration.nodes.prepare_tool_input import (
     prepare_tool_input,
 )
-from onyx.agents.agent_search.orchestration.nodes.tool_call import tool_call
+from onyx.agents.agent_search.orchestration.nodes.use_tool_response import (
+    basic_use_tool_response,
+)
 from onyx.utils.logger import setup_logger
 
 logger = setup_logger()
@@ -33,13 +33,13 @@ def basic_graph_builder() -> StateGraph:
     )
 
     graph.add_node(
-        node="llm_tool_choice",
-        action=llm_tool_choice,
+        node="choose_tool",
+        action=choose_tool,
     )
 
     graph.add_node(
-        node="tool_call",
-        action=tool_call,
+        node="call_tool",
+        action=call_tool,
     )
 
     graph.add_node(
@@ -51,12 +51,12 @@ def basic_graph_builder() -> StateGraph:
 
     graph.add_edge(start_key=START, end_key="prepare_tool_input")
 
-    graph.add_edge(start_key="prepare_tool_input", end_key="llm_tool_choice")
+    graph.add_edge(start_key="prepare_tool_input", end_key="choose_tool")
 
-    graph.add_conditional_edges("llm_tool_choice", should_continue, ["tool_call", END])
+    graph.add_conditional_edges("choose_tool", should_continue, ["call_tool", END])
 
     graph.add_edge(
-        start_key="tool_call",
+        start_key="call_tool",
         end_key="basic_use_tool_response",
     )
 
@@ -73,7 +73,7 @@ def should_continue(state: BasicState) -> str:
         # If there are no tool calls, basic graph already streamed the answer
         END
         if state.tool_choice is None
-        else "tool_call"
+        else "call_tool"
     )
 
 

backend/onyx/agents/agent_search/deep_search/initial/generate_individual_sub_answer/nodes/check_sub_answer.py

@@ -31,12 +31,14 @@ from onyx.agents.agent_search.shared_graph_utils.utils import (
     get_langgraph_node_log_string,
 )
 from onyx.agents.agent_search.shared_graph_utils.utils import parse_question_id
-from onyx.configs.agent_configs import AGENT_TIMEOUT_OVERRIDE_LLM_SUBANSWER_CHECK
+from onyx.configs.agent_configs import AGENT_TIMEOUT_CONNECT_LLM_SUBANSWER_CHECK
+from onyx.configs.agent_configs import AGENT_TIMEOUT_LLM_SUBANSWER_CHECK
 from onyx.llm.chat_llm import LLMRateLimitError
 from onyx.llm.chat_llm import LLMTimeoutError
 from onyx.prompts.agent_search import SUB_ANSWER_CHECK_PROMPT
 from onyx.prompts.agent_search import UNKNOWN_ANSWER
 from onyx.utils.logger import setup_logger
+from onyx.utils.threadpool_concurrency import run_with_timeout
 from onyx.utils.timing import log_function_time
 
 logger = setup_logger()
@@ -85,9 +87,11 @@ def check_sub_answer(
     agent_error: AgentErrorLog | None = None
     response: BaseMessage | None = None
     try:
-        response = fast_llm.invoke(
+        response = run_with_timeout(
+            AGENT_TIMEOUT_LLM_SUBANSWER_CHECK,
+            fast_llm.invoke,
             prompt=msg,
-            timeout_override=AGENT_TIMEOUT_OVERRIDE_LLM_SUBANSWER_CHECK,
+            timeout_override=AGENT_TIMEOUT_CONNECT_LLM_SUBANSWER_CHECK,
         )
 
         quality_str: str = cast(str, response.content)
@@ -96,7 +100,7 @@ def check_sub_answer(
         )
         log_result = f"Answer quality: {quality_str}"
 
-    except LLMTimeoutError:
+    except (LLMTimeoutError, TimeoutError):
         agent_error = AgentErrorLog(
             error_type=AgentLLMErrorType.TIMEOUT,
             error_message=AGENT_LLM_TIMEOUT_MESSAGE,

backend/onyx/agents/agent_search/deep_search/initial/generate_individual_sub_answer/nodes/generate_sub_answer.py

@@ -1,5 +1,4 @@
 from datetime import datetime
-from typing import Any
 from typing import cast
 
 from langchain_core.messages import merge_message_runs
@@ -47,11 +46,13 @@ from onyx.chat.models import StreamStopInfo
 from onyx.chat.models import StreamStopReason
 from onyx.chat.models import StreamType
 from onyx.configs.agent_configs import AGENT_MAX_ANSWER_CONTEXT_DOCS
-from onyx.configs.agent_configs import AGENT_TIMEOUT_OVERRIDE_LLM_SUBANSWER_GENERATION
+from onyx.configs.agent_configs import AGENT_TIMEOUT_CONNECT_LLM_SUBANSWER_GENERATION
+from onyx.configs.agent_configs import AGENT_TIMEOUT_LLM_SUBANSWER_GENERATION
 from onyx.llm.chat_llm import LLMRateLimitError
 from onyx.llm.chat_llm import LLMTimeoutError
 from onyx.prompts.agent_search import NO_RECOVERED_DOCS
 from onyx.utils.logger import setup_logger
+from onyx.utils.threadpool_concurrency import run_with_timeout
 from onyx.utils.timing import log_function_time
 
 logger = setup_logger()
@@ -110,15 +111,14 @@ def generate_sub_answer(
             config=fast_llm.config,
         )
 
-        response: list[str | list[str | dict[str, Any]]] = []
         dispatch_timings: list[float] = []
-
         agent_error: AgentErrorLog | None = None
+        response: list[str] = []
 
-        try:
+        def stream_sub_answer() -> list[str]:
             for message in fast_llm.stream(
                 prompt=msg,
-                timeout_override=AGENT_TIMEOUT_OVERRIDE_LLM_SUBANSWER_GENERATION,
+                timeout_override=AGENT_TIMEOUT_CONNECT_LLM_SUBANSWER_GENERATION,
             ):
                 # TODO: in principle, the answer here COULD contain images, but we don't support that yet
                 content = message.content
@@ -142,8 +142,15 @@ def generate_sub_answer(
                     (end_stream_token - start_stream_token).microseconds
                 )
                 response.append(content)
+            return response
 
-        except LLMTimeoutError:
+        try:
+            response = run_with_timeout(
+                AGENT_TIMEOUT_LLM_SUBANSWER_GENERATION,
+                stream_sub_answer,
+            )
+
+        except (LLMTimeoutError, TimeoutError):
             agent_error = AgentErrorLog(
                 error_type=AgentLLMErrorType.TIMEOUT,
                 error_message=AGENT_LLM_TIMEOUT_MESSAGE,

backend/onyx/agents/agent_search/deep_search/initial/generate_initial_answer/nodes/generate_initial_answer.py

@@ -1,5 +1,4 @@
 from datetime import datetime
-from typing import Any
 from typing import cast
 
 from langchain_core.messages import HumanMessage
@@ -60,11 +59,15 @@ from onyx.agents.agent_search.shared_graph_utils.utils import write_custom_event
 from onyx.chat.models import AgentAnswerPiece
 from onyx.chat.models import ExtendedToolResponse
 from onyx.chat.models import StreamingError
+from onyx.configs.agent_configs import AGENT_ANSWER_GENERATION_BY_FAST_LLM
 from onyx.configs.agent_configs import AGENT_MAX_ANSWER_CONTEXT_DOCS
 from onyx.configs.agent_configs import AGENT_MAX_STREAMED_DOCS_FOR_INITIAL_ANSWER
 from onyx.configs.agent_configs import AGENT_MIN_ORIG_QUESTION_DOCS
 from onyx.configs.agent_configs import (
-    AGENT_TIMEOUT_OVERRIDE_LLM_INITIAL_ANSWER_GENERATION,
+    AGENT_TIMEOUT_CONNECT_LLM_INITIAL_ANSWER_GENERATION,
+)
+from onyx.configs.agent_configs import (
+    AGENT_TIMEOUT_LLM_INITIAL_ANSWER_GENERATION,
 )
 from onyx.llm.chat_llm import LLMRateLimitError
 from onyx.llm.chat_llm import LLMTimeoutError
@@ -77,6 +80,7 @@ from onyx.prompts.agent_search import (
 )
 from onyx.prompts.agent_search import UNKNOWN_ANSWER
 from onyx.tools.tool_implementations.search.search_tool import yield_search_responses
+from onyx.utils.threadpool_concurrency import run_with_timeout
 from onyx.utils.timing import log_function_time
 
 _llm_node_error_strings = LLMNodeErrorStrings(
@@ -230,7 +234,11 @@ def generate_initial_answer(
 
         sub_questions = all_sub_questions  # Replace the original assignment
 
-        model = graph_config.tooling.fast_llm
+        model = (
+            graph_config.tooling.fast_llm
+            if AGENT_ANSWER_GENERATION_BY_FAST_LLM
+            else graph_config.tooling.primary_llm
+        )
 
         doc_context = format_docs(answer_generation_documents.context_documents)
         doc_context = trim_prompt_piece(
@@ -260,15 +268,16 @@ def generate_initial_answer(
             )
         ]
 
-        streamed_tokens: list[str | list[str | dict[str, Any]]] = [""]
+        streamed_tokens: list[str] = [""]
         dispatch_timings: list[float] = []
 
         agent_error: AgentErrorLog | None = None
 
-        try:
+        def stream_initial_answer() -> list[str]:
+            response: list[str] = []
             for message in model.stream(
                 msg,
-                timeout_override=AGENT_TIMEOUT_OVERRIDE_LLM_INITIAL_ANSWER_GENERATION,
+                timeout_override=AGENT_TIMEOUT_CONNECT_LLM_INITIAL_ANSWER_GENERATION,
             ):
                 # TODO: in principle, the answer here COULD contain images, but we don't support that yet
                 content = message.content
@@ -292,9 +301,16 @@ def generate_initial_answer(
                 dispatch_timings.append(
                     (end_stream_token - start_stream_token).microseconds
                 )
-                streamed_tokens.append(content)
+                response.append(content)
+            return response
 
-        except LLMTimeoutError:
+        try:
+            streamed_tokens = run_with_timeout(
+                AGENT_TIMEOUT_LLM_INITIAL_ANSWER_GENERATION,
+                stream_initial_answer,
+            )
+
+        except (LLMTimeoutError, TimeoutError):
             agent_error = AgentErrorLog(
                 error_type=AgentLLMErrorType.TIMEOUT,
                 error_message=AGENT_LLM_TIMEOUT_MESSAGE,

backend/onyx/agents/agent_search/deep_search/initial/generate_sub_answers/nodes/decompose_orig_question.py

@@ -36,7 +36,10 @@ from onyx.chat.models import StreamType
 from onyx.chat.models import SubQuestionPiece
 from onyx.configs.agent_configs import AGENT_NUM_DOCS_FOR_DECOMPOSITION
 from onyx.configs.agent_configs import (
-    AGENT_TIMEOUT_OVERRIDE_LLM_SUBQUESTION_GENERATION,
+    AGENT_TIMEOUT_CONNECT_LLM_SUBQUESTION_GENERATION,
+)
+from onyx.configs.agent_configs import (
+    AGENT_TIMEOUT_LLM_SUBQUESTION_GENERATION,
 )
 from onyx.llm.chat_llm import LLMRateLimitError
 from onyx.llm.chat_llm import LLMTimeoutError
@@ -47,6 +50,7 @@ from onyx.prompts.agent_search import (
     INITIAL_QUESTION_DECOMPOSITION_PROMPT_ASSUMING_REFINEMENT,
 )
 from onyx.utils.logger import setup_logger
+from onyx.utils.threadpool_concurrency import run_with_timeout
 from onyx.utils.timing import log_function_time
 
 logger = setup_logger()
@@ -131,10 +135,12 @@ def decompose_orig_question(
     streamed_tokens: list[BaseMessage_Content] = []
 
     try:
-        streamed_tokens = dispatch_separated(
+        streamed_tokens = run_with_timeout(
+            AGENT_TIMEOUT_LLM_SUBQUESTION_GENERATION,
+            dispatch_separated,
             model.stream(
                 msg,
-                timeout_override=AGENT_TIMEOUT_OVERRIDE_LLM_SUBQUESTION_GENERATION,
+                timeout_override=AGENT_TIMEOUT_CONNECT_LLM_SUBQUESTION_GENERATION,
             ),
             dispatch_subquestion(0, writer),
             sep_callback=dispatch_subquestion_sep(0, writer),
@@ -154,7 +160,7 @@ def decompose_orig_question(
         )
         write_custom_event("stream_finished", stop_event, writer)
 
-    except LLMTimeoutError as e:
+    except (LLMTimeoutError, TimeoutError) as e:
         logger.error("LLM Timeout Error - decompose orig question")
         raise e  # fail loudly on this critical step
     except LLMRateLimitError as e:

backend/onyx/agents/agent_search/deep_search/main/edges.py

@@ -25,7 +25,7 @@ logger = setup_logger()
 
 def route_initial_tool_choice(
     state: MainState, config: RunnableConfig
-) -> Literal["tool_call", "start_agent_search", "logging_node"]:
+) -> Literal["call_tool", "start_agent_search", "logging_node"]:
     """
     LangGraph edge to route to agent search.
     """
@@ -38,7 +38,7 @@ def route_initial_tool_choice(
         ):
             return "start_agent_search"
         else:
-            return "tool_call"
+            return "call_tool"
     else:
         return "logging_node"
 

backend/onyx/agents/agent_search/deep_search/main/graph_builder.py

@@ -43,14 +43,14 @@ from onyx.agents.agent_search.deep_search.main.states import MainState
 from onyx.agents.agent_search.deep_search.refinement.consolidate_sub_answers.graph_builder import (
     answer_refined_query_graph_builder,
 )
-from onyx.agents.agent_search.orchestration.nodes.basic_use_tool_response import (
-    basic_use_tool_response,
-)
-from onyx.agents.agent_search.orchestration.nodes.llm_tool_choice import llm_tool_choice
+from onyx.agents.agent_search.orchestration.nodes.call_tool import call_tool
+from onyx.agents.agent_search.orchestration.nodes.choose_tool import choose_tool
 from onyx.agents.agent_search.orchestration.nodes.prepare_tool_input import (
     prepare_tool_input,
 )
-from onyx.agents.agent_search.orchestration.nodes.tool_call import tool_call
+from onyx.agents.agent_search.orchestration.nodes.use_tool_response import (
+    basic_use_tool_response,
+)
 from onyx.agents.agent_search.shared_graph_utils.utils import get_test_config
 from onyx.utils.logger import setup_logger
 
@@ -77,13 +77,13 @@ def main_graph_builder(test_mode: bool = False) -> StateGraph:
     # Choose the initial tool
     graph.add_node(
         node="initial_tool_choice",
-        action=llm_tool_choice,
+        action=choose_tool,
     )
 
     # Call the tool, if required
     graph.add_node(
-        node="tool_call",
-        action=tool_call,
+        node="call_tool",
+        action=call_tool,
     )
 
     # Use the tool response
@@ -168,11 +168,11 @@ def main_graph_builder(test_mode: bool = False) -> StateGraph:
     graph.add_conditional_edges(
         "initial_tool_choice",
         route_initial_tool_choice,
-        ["tool_call", "start_agent_search", "logging_node"],
+        ["call_tool", "start_agent_search", "logging_node"],
     )
 
     graph.add_edge(
-        start_key="tool_call",
+        start_key="call_tool",
         end_key="basic_use_tool_response",
     )
     graph.add_edge(

backend/onyx/agents/agent_search/deep_search/main/nodes/compare_answers.py

@@ -33,13 +33,15 @@ from onyx.agents.agent_search.shared_graph_utils.utils import (
 )
 from onyx.agents.agent_search.shared_graph_utils.utils import write_custom_event
 from onyx.chat.models import RefinedAnswerImprovement
-from onyx.configs.agent_configs import AGENT_TIMEOUT_OVERRIDE_LLM_COMPARE_ANSWERS
+from onyx.configs.agent_configs import AGENT_TIMEOUT_CONNECT_LLM_COMPARE_ANSWERS
+from onyx.configs.agent_configs import AGENT_TIMEOUT_LLM_COMPARE_ANSWERS
 from onyx.llm.chat_llm import LLMRateLimitError
 from onyx.llm.chat_llm import LLMTimeoutError
 from onyx.prompts.agent_search import (
     INITIAL_REFINED_ANSWER_COMPARISON_PROMPT,
 )
 from onyx.utils.logger import setup_logger
+from onyx.utils.threadpool_concurrency import run_with_timeout
 from onyx.utils.timing import log_function_time
 
 logger = setup_logger()
@@ -105,11 +107,14 @@ def compare_answers(
     refined_answer_improvement: bool | None = None
     # no need to stream this
     try:
-        resp = model.invoke(
-            msg, timeout_override=AGENT_TIMEOUT_OVERRIDE_LLM_COMPARE_ANSWERS
+        resp = run_with_timeout(
+            AGENT_TIMEOUT_LLM_COMPARE_ANSWERS,
+            model.invoke,
+            prompt=msg,
+            timeout_override=AGENT_TIMEOUT_CONNECT_LLM_COMPARE_ANSWERS,
         )
 
-    except LLMTimeoutError:
+    except (LLMTimeoutError, TimeoutError):
         agent_error = AgentErrorLog(
             error_type=AgentLLMErrorType.TIMEOUT,
             error_message=AGENT_LLM_TIMEOUT_MESSAGE,

backend/onyx/agents/agent_search/deep_search/main/nodes/create_refined_sub_questions.py

@@ -44,7 +44,10 @@ from onyx.agents.agent_search.shared_graph_utils.utils import make_question_id
 from onyx.agents.agent_search.shared_graph_utils.utils import write_custom_event
 from onyx.chat.models import StreamingError
 from onyx.configs.agent_configs import (
-    AGENT_TIMEOUT_OVERRIDE_LLM_REFINED_SUBQUESTION_GENERATION,
+    AGENT_TIMEOUT_CONNECT_LLM_REFINED_SUBQUESTION_GENERATION,
+)
+from onyx.configs.agent_configs import (
+    AGENT_TIMEOUT_LLM_REFINED_SUBQUESTION_GENERATION,
 )
 from onyx.llm.chat_llm import LLMRateLimitError
 from onyx.llm.chat_llm import LLMTimeoutError
@@ -53,6 +56,7 @@ from onyx.prompts.agent_search import (
 )
 from onyx.tools.models import ToolCallKickoff
 from onyx.utils.logger import setup_logger
+from onyx.utils.threadpool_concurrency import run_with_timeout
 from onyx.utils.timing import log_function_time
 
 logger = setup_logger()
@@ -134,15 +138,17 @@ def create_refined_sub_questions(
     agent_error: AgentErrorLog | None = None
     streamed_tokens: list[BaseMessage_Content] = []
     try:
-        streamed_tokens = dispatch_separated(
+        streamed_tokens = run_with_timeout(
+            AGENT_TIMEOUT_LLM_REFINED_SUBQUESTION_GENERATION,
+            dispatch_separated,
             model.stream(
                 msg,
-                timeout_override=AGENT_TIMEOUT_OVERRIDE_LLM_REFINED_SUBQUESTION_GENERATION,
+                timeout_override=AGENT_TIMEOUT_CONNECT_LLM_REFINED_SUBQUESTION_GENERATION,
             ),
             dispatch_subquestion(1, writer),
             sep_callback=dispatch_subquestion_sep(1, writer),
         )
-    except LLMTimeoutError:
+    except (LLMTimeoutError, TimeoutError):
         agent_error = AgentErrorLog(
             error_type=AgentLLMErrorType.TIMEOUT,
             error_message=AGENT_LLM_TIMEOUT_MESSAGE,

backend/onyx/agents/agent_search/deep_search/main/nodes/extract_entities_terms.py

@@ -22,11 +22,17 @@ from onyx.agents.agent_search.shared_graph_utils.utils import (
     get_langgraph_node_log_string,
 )
 from onyx.configs.agent_configs import (
-    AGENT_TIMEOUT_OVERRIDE_LLM_ENTITY_TERM_EXTRACTION,
+    AGENT_TIMEOUT_CONNECT_LLM_ENTITY_TERM_EXTRACTION,
+)
+from onyx.configs.agent_configs import (
+    AGENT_TIMEOUT_LLM_ENTITY_TERM_EXTRACTION,
 )
 from onyx.configs.constants import NUM_EXPLORATORY_DOCS
+from onyx.llm.chat_llm import LLMRateLimitError
+from onyx.llm.chat_llm import LLMTimeoutError
 from onyx.prompts.agent_search import ENTITY_TERM_EXTRACTION_PROMPT
 from onyx.prompts.agent_search import ENTITY_TERM_EXTRACTION_PROMPT_JSON_EXAMPLE
+from onyx.utils.threadpool_concurrency import run_with_timeout
 from onyx.utils.timing import log_function_time
 
 
@@ -84,30 +90,42 @@ def extract_entities_terms(
     ]
     fast_llm = graph_config.tooling.fast_llm
     # Grader
-    llm_response = fast_llm.invoke(
-        prompt=msg,
-        timeout_override=AGENT_TIMEOUT_OVERRIDE_LLM_ENTITY_TERM_EXTRACTION,
-    )
-
-    cleaned_response = (
-        str(llm_response.content).replace("```json\n", "").replace("\n```", "")
-    )
-    first_bracket = cleaned_response.find("{")
-    last_bracket = cleaned_response.rfind("}")
-    cleaned_response = cleaned_response[first_bracket : last_bracket + 1]
-
     try:
-        entity_extraction_result = EntityExtractionResult.model_validate_json(
-            cleaned_response
+        llm_response = run_with_timeout(
+            AGENT_TIMEOUT_LLM_ENTITY_TERM_EXTRACTION,
+            fast_llm.invoke,
+            prompt=msg,
+            timeout_override=AGENT_TIMEOUT_CONNECT_LLM_ENTITY_TERM_EXTRACTION,
         )
-    except ValueError:
-        logger.error("Failed to parse LLM response as JSON in Entity-Term Extraction")
+
+        cleaned_response = (
+            str(llm_response.content).replace("```json\n", "").replace("\n```", "")
+        )
+        first_bracket = cleaned_response.find("{")
+        last_bracket = cleaned_response.rfind("}")
+        cleaned_response = cleaned_response[first_bracket : last_bracket + 1]
+
+        try:
+            entity_extraction_result = EntityExtractionResult.model_validate_json(
+                cleaned_response
+            )
+        except ValueError:
+            logger.error(
+                "Failed to parse LLM response as JSON in Entity-Term Extraction"
+            )
+            entity_extraction_result = EntityExtractionResult(
+                retrieved_entities_relationships=EntityRelationshipTermExtraction(),
+            )
+    except (LLMTimeoutError, TimeoutError):
+        logger.error("LLM Timeout Error - extract entities terms")
         entity_extraction_result = EntityExtractionResult(
-            retrieved_entities_relationships=EntityRelationshipTermExtraction(
-                entities=[],
-                relationships=[],
-                terms=[],
-            ),
+            retrieved_entities_relationships=EntityRelationshipTermExtraction(),
+        )
+
+    except LLMRateLimitError:
+        logger.error("LLM Rate Limit Error - extract entities terms")
+        entity_extraction_result = EntityExtractionResult(
+            retrieved_entities_relationships=EntityRelationshipTermExtraction(),
         )
 
     return EntityTermExtractionUpdate(

backend/onyx/agents/agent_search/deep_search/main/nodes/generate_validate_refined_answer.py

@@ -1,5 +1,4 @@
 from datetime import datetime
-from typing import Any
 from typing import cast
 
 from langchain_core.messages import HumanMessage
@@ -66,14 +65,21 @@ from onyx.agents.agent_search.shared_graph_utils.utils import write_custom_event
 from onyx.chat.models import AgentAnswerPiece
 from onyx.chat.models import ExtendedToolResponse
 from onyx.chat.models import StreamingError
+from onyx.configs.agent_configs import AGENT_ANSWER_GENERATION_BY_FAST_LLM
 from onyx.configs.agent_configs import AGENT_MAX_ANSWER_CONTEXT_DOCS
 from onyx.configs.agent_configs import AGENT_MAX_STREAMED_DOCS_FOR_REFINED_ANSWER
 from onyx.configs.agent_configs import AGENT_MIN_ORIG_QUESTION_DOCS
 from onyx.configs.agent_configs import (
-    AGENT_TIMEOUT_OVERRIDE_LLM_REFINED_ANSWER_GENERATION,
+    AGENT_TIMEOUT_CONNECT_LLM_REFINED_ANSWER_GENERATION,
 )
 from onyx.configs.agent_configs import (
-    AGENT_TIMEOUT_OVERRIDE_LLM_REFINED_ANSWER_VALIDATION,
+    AGENT_TIMEOUT_CONNECT_LLM_REFINED_ANSWER_VALIDATION,
+)
+from onyx.configs.agent_configs import (
+    AGENT_TIMEOUT_LLM_REFINED_ANSWER_GENERATION,
+)
+from onyx.configs.agent_configs import (
+    AGENT_TIMEOUT_LLM_REFINED_ANSWER_VALIDATION,
 )
 from onyx.llm.chat_llm import LLMRateLimitError
 from onyx.llm.chat_llm import LLMTimeoutError
@@ -92,6 +98,7 @@ from onyx.prompts.agent_search import (
 from onyx.prompts.agent_search import UNKNOWN_ANSWER
 from onyx.tools.tool_implementations.search.search_tool import yield_search_responses
 from onyx.utils.logger import setup_logger
+from onyx.utils.threadpool_concurrency import run_with_timeout
 from onyx.utils.timing import log_function_time
 
 logger = setup_logger()
@@ -253,7 +260,12 @@ def generate_validate_refined_answer(
         else REFINED_ANSWER_PROMPT_WO_SUB_QUESTIONS
     )
 
-    model = graph_config.tooling.fast_llm
+    model = (
+        graph_config.tooling.fast_llm
+        if AGENT_ANSWER_GENERATION_BY_FAST_LLM
+        else graph_config.tooling.primary_llm
+    )
+
     relevant_docs_str = format_docs(answer_generation_documents.context_documents)
     relevant_docs_str = trim_prompt_piece(
         model.config,
@@ -284,13 +296,13 @@ def generate_validate_refined_answer(
         )
     ]
 
-    streamed_tokens: list[str | list[str | dict[str, Any]]] = [""]
+    streamed_tokens: list[str] = [""]
     dispatch_timings: list[float] = []
     agent_error: AgentErrorLog | None = None
 
-    try:
+    def stream_refined_answer() -> list[str]:
         for message in model.stream(
-            msg, timeout_override=AGENT_TIMEOUT_OVERRIDE_LLM_REFINED_ANSWER_GENERATION
+            msg, timeout_override=AGENT_TIMEOUT_CONNECT_LLM_REFINED_ANSWER_GENERATION
         ):
             # TODO: in principle, the answer here COULD contain images, but we don't support that yet
             content = message.content
@@ -315,8 +327,15 @@ def generate_validate_refined_answer(
                 (end_stream_token - start_stream_token).microseconds
             )
             streamed_tokens.append(content)
+        return streamed_tokens
 
-    except LLMTimeoutError:
+    try:
+        streamed_tokens = run_with_timeout(
+            AGENT_TIMEOUT_LLM_REFINED_ANSWER_GENERATION,
+            stream_refined_answer,
+        )
+
+    except (LLMTimeoutError, TimeoutError):
         agent_error = AgentErrorLog(
             error_type=AgentLLMErrorType.TIMEOUT,
             error_message=AGENT_LLM_TIMEOUT_MESSAGE,
@@ -383,16 +402,20 @@ def generate_validate_refined_answer(
         )
     ]
 
+    validation_model = graph_config.tooling.fast_llm
     try:
-        validation_response = model.invoke(
-            msg, timeout_override=AGENT_TIMEOUT_OVERRIDE_LLM_REFINED_ANSWER_VALIDATION
+        validation_response = run_with_timeout(
+            AGENT_TIMEOUT_LLM_REFINED_ANSWER_VALIDATION,
+            validation_model.invoke,
+            prompt=msg,
+            timeout_override=AGENT_TIMEOUT_CONNECT_LLM_REFINED_ANSWER_VALIDATION,
         )
         refined_answer_quality = binary_string_test_after_answer_separator(
             text=cast(str, validation_response.content),
             positive_value=AGENT_POSITIVE_VALUE_STR,
             separator=AGENT_ANSWER_SEPARATOR,
         )
-    except LLMTimeoutError:
+    except (LLMTimeoutError, TimeoutError):
         refined_answer_quality = True
         logger.error("LLM Timeout Error - validate refined answer")
 

backend/onyx/agents/agent_search/deep_search/shared/expanded_retrieval/nodes/expand_queries.py

@@ -34,14 +34,16 @@ from onyx.agents.agent_search.shared_graph_utils.utils import (
 )
 from onyx.agents.agent_search.shared_graph_utils.utils import parse_question_id
 from onyx.configs.agent_configs import (
-    AGENT_TIMEOUT_OVERRIDE_LLM_QUERY_REWRITING_GENERATION,
+    AGENT_TIMEOUT_CONNECT_LLM_QUERY_REWRITING_GENERATION,
 )
+from onyx.configs.agent_configs import AGENT_TIMEOUT_LLM_QUERY_REWRITING_GENERATION
 from onyx.llm.chat_llm import LLMRateLimitError
 from onyx.llm.chat_llm import LLMTimeoutError
 from onyx.prompts.agent_search import (
     QUERY_REWRITING_PROMPT,
 )
 from onyx.utils.logger import setup_logger
+from onyx.utils.threadpool_concurrency import run_with_timeout
 from onyx.utils.timing import log_function_time
 
 logger = setup_logger()
@@ -69,7 +71,7 @@ def expand_queries(
     node_start_time = datetime.now()
     question = state.question
 
-    llm = graph_config.tooling.fast_llm
+    model = graph_config.tooling.fast_llm
     sub_question_id = state.sub_question_id
     if sub_question_id is None:
         level, question_num = 0, 0
@@ -88,10 +90,12 @@ def expand_queries(
     rewritten_queries = []
 
     try:
-        llm_response_list = dispatch_separated(
-            llm.stream(
+        llm_response_list = run_with_timeout(
+            AGENT_TIMEOUT_LLM_QUERY_REWRITING_GENERATION,
+            dispatch_separated,
+            model.stream(
                 prompt=msg,
-                timeout_override=AGENT_TIMEOUT_OVERRIDE_LLM_QUERY_REWRITING_GENERATION,
+                timeout_override=AGENT_TIMEOUT_CONNECT_LLM_QUERY_REWRITING_GENERATION,
             ),
             dispatch_subquery(level, question_num, writer),
         )
@@ -101,7 +105,7 @@ def expand_queries(
         rewritten_queries = llm_response.split("\n")
         log_result = f"Number of expanded queries: {len(rewritten_queries)}"
 
-    except LLMTimeoutError:
+    except (LLMTimeoutError, TimeoutError):
         agent_error = AgentErrorLog(
             error_type=AgentLLMErrorType.TIMEOUT,
             error_message=AGENT_LLM_TIMEOUT_MESSAGE,

backend/onyx/agents/agent_search/deep_search/shared/expanded_retrieval/nodes/rerank_documents.py

@@ -55,6 +55,7 @@ def rerank_documents(
 
     # Note that these are passed in values from the API and are overrides which are typically None
     rerank_settings = graph_config.inputs.search_request.rerank_settings
+    allow_agent_reranking = graph_config.behavior.allow_agent_reranking
 
     if rerank_settings is None:
         with get_session_context_manager() as db_session:
@@ -62,23 +63,31 @@ def rerank_documents(
             if not search_settings.disable_rerank_for_streaming:
                 rerank_settings = RerankingDetails.from_db_model(search_settings)
 
+    # Initial default: no reranking. Will be overwritten below if reranking is warranted
+    reranked_documents = verified_documents
+
     if should_rerank(rerank_settings) and len(verified_documents) > 0:
         if len(verified_documents) > 1:
-            reranked_documents = rerank_sections(
-                query_str=question,
-                # if runnable, then rerank_settings is not None
-                rerank_settings=cast(RerankingDetails, rerank_settings),
-                sections_to_rerank=verified_documents,
-            )
+            if not allow_agent_reranking:
+                logger.info("Use of local rerank model without GPU, skipping reranking")
+            # No reranking, stay with verified_documents as default
+
+            else:
+                # Reranking is warranted, use the rerank_sections functon
+                reranked_documents = rerank_sections(
+                    query_str=question,
+                    # if runnable, then rerank_settings is not None
+                    rerank_settings=cast(RerankingDetails, rerank_settings),
+                    sections_to_rerank=verified_documents,
+                )
         else:
             logger.warning(
                 f"{len(verified_documents)} verified document(s) found, skipping reranking"
             )
-            reranked_documents = verified_documents
+            # No reranking, stay with verified_documents as default
     else:
         logger.warning("No reranking settings found, using unranked documents")
-        reranked_documents = verified_documents
-
+        # No reranking, stay with verified_documents as default
     if AGENT_RERANKING_STATS:
         fit_scores = get_fit_scores(verified_documents, reranked_documents)
     else:

backend/onyx/agents/agent_search/deep_search/shared/expanded_retrieval/nodes/verify_documents.py

@@ -25,13 +25,15 @@ from onyx.agents.agent_search.shared_graph_utils.models import LLMNodeErrorStrin
 from onyx.agents.agent_search.shared_graph_utils.utils import (
     get_langgraph_node_log_string,
 )
-from onyx.configs.agent_configs import AGENT_TIMEOUT_OVERRIDE_LLM_DOCUMENT_VERIFICATION
+from onyx.configs.agent_configs import AGENT_TIMEOUT_CONNECT_LLM_DOCUMENT_VERIFICATION
+from onyx.configs.agent_configs import AGENT_TIMEOUT_LLM_DOCUMENT_VERIFICATION
 from onyx.llm.chat_llm import LLMRateLimitError
 from onyx.llm.chat_llm import LLMTimeoutError
 from onyx.prompts.agent_search import (
     DOCUMENT_VERIFICATION_PROMPT,
 )
 from onyx.utils.logger import setup_logger
+from onyx.utils.threadpool_concurrency import run_with_timeout
 from onyx.utils.timing import log_function_time
 
 logger = setup_logger()
@@ -86,8 +88,11 @@ def verify_documents(
     ]  # default is to treat document as relevant
 
     try:
-        response = fast_llm.invoke(
-            msg, timeout_override=AGENT_TIMEOUT_OVERRIDE_LLM_DOCUMENT_VERIFICATION
+        response = run_with_timeout(
+            AGENT_TIMEOUT_LLM_DOCUMENT_VERIFICATION,
+            fast_llm.invoke,
+            prompt=msg,
+            timeout_override=AGENT_TIMEOUT_CONNECT_LLM_DOCUMENT_VERIFICATION,
         )
 
         assert isinstance(response.content, str)
@@ -96,7 +101,7 @@ def verify_documents(
         ):
             verified_documents = []
 
-    except LLMTimeoutError:
+    except (LLMTimeoutError, TimeoutError):
         # In this case, we decide to continue and don't raise an error, as
         # little harm in letting some docs through that are less relevant.
         logger.error("LLM Timeout Error - verify documents")

backend/onyx/agents/agent_search/models.py

@@ -67,6 +67,7 @@ class GraphSearchConfig(BaseModel):
     # Whether to allow creation of refinement questions (and entity extraction, etc.)
     allow_refinement: bool = True
     skip_gen_ai_answer_generation: bool = False
+    allow_agent_reranking: bool = False
 
 
 class GraphConfig(BaseModel):

backend/onyx/agents/agent_search/orchestration/nodes/call_tool.py

@@ -28,7 +28,7 @@ def emit_packet(packet: AnswerPacket, writer: StreamWriter) -> None:
     write_custom_event("basic_response", packet, writer)
 
 
-def tool_call(
+def call_tool(
     state: ToolChoiceUpdate,
     config: RunnableConfig,
     writer: StreamWriter = lambda _: None,

backend/onyx/agents/agent_search/orchestration/nodes/choose_tool.py

@@ -25,7 +25,7 @@ logger = setup_logger()
 # and a function that handles extracting the necessary fields
 # from the state and config
 # TODO: fan-out to multiple tool call nodes? Make this configurable?
-def llm_tool_choice(
+def choose_tool(
     state: ToolChoiceState,
     config: RunnableConfig,
     writer: StreamWriter = lambda _: None,
@@ -98,8 +98,16 @@ def llm_tool_choice(
         # For tool calling LLMs, we want to insert the task prompt as part of this flow, this is because the LLM
         # may choose to not call any tools and just generate the answer, in which case the task prompt is needed.
         prompt=built_prompt,
-        tools=[tool.tool_definition() for tool in tools] or None,
-        tool_choice=("required" if tools and force_use_tool.force_use else None),
+        tools=(
+            [tool.tool_definition() for tool in tools] or None
+            if using_tool_calling_llm
+            else None
+        ),
+        tool_choice=(
+            "required"
+            if tools and force_use_tool.force_use and using_tool_calling_llm
+            else None
+        ),
         structured_response_format=structured_response_format,
     )
 

backend/onyx/agents/agent_search/shared_graph_utils/utils.py

@@ -43,8 +43,9 @@ from onyx.chat.models import StreamStopReason
 from onyx.chat.models import StreamType
 from onyx.chat.prompt_builder.answer_prompt_builder import AnswerPromptBuilder
 from onyx.configs.agent_configs import (
-    AGENT_TIMEOUT_OVERRIDE_LLM_HISTORY_SUMMARY_GENERATION,
+    AGENT_TIMEOUT_CONNECT_LLM_HISTORY_SUMMARY_GENERATION,
 )
+from onyx.configs.agent_configs import AGENT_TIMEOUT_LLM_HISTORY_SUMMARY_GENERATION
 from onyx.configs.chat_configs import CHAT_TARGET_CHUNK_PERCENTAGE
 from onyx.configs.chat_configs import MAX_CHUNKS_FED_TO_CHAT
 from onyx.configs.constants import DEFAULT_PERSONA_ID
@@ -80,6 +81,7 @@ from onyx.tools.tool_implementations.search.search_tool import SearchResponseSum
 from onyx.tools.tool_implementations.search.search_tool import SearchTool
 from onyx.tools.utils import explicit_tool_calling_supported
 from onyx.utils.logger import setup_logger
+from onyx.utils.threadpool_concurrency import run_with_timeout
 
 logger = setup_logger()
 
@@ -395,11 +397,13 @@ def summarize_history(
     )
 
     try:
-        history_response = llm.invoke(
+        history_response = run_with_timeout(
+            AGENT_TIMEOUT_LLM_HISTORY_SUMMARY_GENERATION,
+            llm.invoke,
             history_context_prompt,
-            timeout_override=AGENT_TIMEOUT_OVERRIDE_LLM_HISTORY_SUMMARY_GENERATION,
+            timeout_override=AGENT_TIMEOUT_CONNECT_LLM_HISTORY_SUMMARY_GENERATION,
         )
-    except LLMTimeoutError:
+    except (LLMTimeoutError, TimeoutError):
         logger.error("LLM Timeout Error - summarize history")
         return (
             history  # this is what is done at this point anyway, so we default to this

backend/onyx/auth/api_key.py

@@ -10,6 +10,7 @@ from pydantic import BaseModel
 
 from onyx.auth.schemas import UserRole
 from onyx.configs.app_configs import API_KEY_HASH_ROUNDS
+from shared_configs.configs import MULTI_TENANT
 
 
 _API_KEY_HEADER_NAME = "Authorization"
@@ -35,8 +36,7 @@ class ApiKeyDescriptor(BaseModel):
 
 
 def generate_api_key(tenant_id: str | None = None) -> str:
-    # For backwards compatibility, if no tenant_id, generate old style key
-    if not tenant_id:
+    if not MULTI_TENANT or not tenant_id:
         return _API_KEY_PREFIX + secrets.token_urlsafe(_API_KEY_LEN)
 
     encoded_tenant = quote(tenant_id)  # URL encode the tenant ID

backend/onyx/auth/email_utils.py

@@ -2,6 +2,8 @@ import smtplib
 from datetime import datetime
 from email.mime.multipart import MIMEMultipart
 from email.mime.text import MIMEText
+from email.utils import formatdate
+from email.utils import make_msgid
 
 from onyx.configs.app_configs import EMAIL_CONFIGURED
 from onyx.configs.app_configs import EMAIL_FROM
@@ -10,8 +12,10 @@ from onyx.configs.app_configs import SMTP_PORT
 from onyx.configs.app_configs import SMTP_SERVER
 from onyx.configs.app_configs import SMTP_USER
 from onyx.configs.app_configs import WEB_DOMAIN
+from onyx.configs.constants import AuthType
 from onyx.configs.constants import TENANT_ID_COOKIE_NAME
 from onyx.db.models import User
+from shared_configs.configs import MULTI_TENANT
 
 HTML_EMAIL_TEMPLATE = """\
 <!DOCTYPE html>
@@ -149,8 +153,9 @@ def send_email(
     msg = MIMEMultipart("alternative")
     msg["Subject"] = subject
     msg["To"] = user_email
-    if mail_from:
-        msg["From"] = mail_from
+    msg["From"] = mail_from
+    msg["Date"] = formatdate(localtime=True)
+    msg["Message-ID"] = make_msgid(domain="onyx.app")
 
     part_text = MIMEText(text_body, "plain")
     part_html = MIMEText(html_body, "html")
@@ -172,7 +177,7 @@ def send_subscription_cancellation_email(user_email: str) -> None:
     subject = "Your Onyx Subscription Has Been Canceled"
     heading = "Subscription Canceled"
     message = (
-        "<p>We’re sorry to see you go.</p>"
+        "<p>We're sorry to see you go.</p>"
         "<p>Your subscription has been canceled and will end on your next billing date.</p>"
         "<p>If you change your mind, you can always come back!</p>"
     )
@@ -187,36 +192,64 @@ def send_subscription_cancellation_email(user_email: str) -> None:
     send_email(user_email, subject, html_content, text_content)
 
 
-def send_user_email_invite(user_email: str, current_user: User) -> None:
+def send_user_email_invite(
+    user_email: str, current_user: User, auth_type: AuthType
+) -> None:
     subject = "Invitation to Join Onyx Organization"
     heading = "You've Been Invited!"
-    message = (
-        f"<p>You have been invited by {current_user.email} to join an organization on Onyx.</p>"
-        "<p>To join the organization, please click the button below to set a password "
-        "or login with Google and complete your registration.</p>"
-    )
+
+    # the exact action taken by the user, and thus the message, depends on the auth type
+    message = f"<p>You have been invited by {current_user.email} to join an organization on Onyx.</p>"
+    if auth_type == AuthType.CLOUD:
+        message += (
+            "<p>To join the organization, please click the button below to set a password "
+            "or login with Google and complete your registration.</p>"
+        )
+    elif auth_type == AuthType.BASIC:
+        message += (
+            "<p>To join the organization, please click the button below to set a password "
+            "and complete your registration.</p>"
+        )
+    elif auth_type == AuthType.GOOGLE_OAUTH:
+        message += (
+            "<p>To join the organization, please click the button below to login with Google "
+            "and complete your registration.</p>"
+        )
+    elif auth_type == AuthType.OIDC or auth_type == AuthType.SAML:
+        message += (
+            "<p>To join the organization, please click the button below to"
+            " complete your registration.</p>"
+        )
+    else:
+        raise ValueError(f"Invalid auth type: {auth_type}")
+
     cta_text = "Join Organization"
     cta_link = f"{WEB_DOMAIN}/auth/signup?email={user_email}"
     html_content = build_html_email(heading, message, cta_text, cta_link)
+
+    # text content is the fallback for clients that don't support HTML
+    # not as critical, so not having special cases for each auth type
     text_content = (
         f"You have been invited by {current_user.email} to join an organization on Onyx.\n"
         "To join the organization, please visit the following link:\n"
         f"{WEB_DOMAIN}/auth/signup?email={user_email}\n"
-        "You'll be asked to set a password or login with Google to complete your registration."
     )
+    if auth_type == AuthType.CLOUD:
+        text_content += "You'll be asked to set a password or login with Google to complete your registration."
+
     send_email(user_email, subject, html_content, text_content)
 
 
 def send_forgot_password_email(
     user_email: str,
     token: str,
+    tenant_id: str,
     mail_from: str = EMAIL_FROM,
-    tenant_id: str | None = None,
 ) -> None:
     # Builds a forgot password email with or without fancy HTML
     subject = "Onyx Forgot Password"
     link = f"{WEB_DOMAIN}/auth/reset-password?token={token}"
-    if tenant_id:
+    if MULTI_TENANT:
         link += f"&{TENANT_ID_COOKIE_NAME}={tenant_id}"
     message = f"<p>Click the following link to reset your password:</p><p>{link}</p>"
     html_content = build_html_email("Reset Your Password", message)

backend/onyx/auth/noauth_user.py

@@ -42,4 +42,5 @@ def fetch_no_auth_user(
         role=UserRole.BASIC if anonymous_user_enabled else UserRole.ADMIN,
         preferences=load_no_auth_user_preferences(store),
         is_anonymous_user=anonymous_user_enabled,
+        password_configured=False,
     )

backend/onyx/auth/users.py

@@ -1,5 +1,7 @@
 import json
+import random
 import secrets
+import string
 import uuid
 from collections.abc import AsyncGenerator
 from datetime import datetime
@@ -86,7 +88,6 @@ from onyx.db.auth import get_user_db
 from onyx.db.auth import SQLAlchemyUserAdminDB
 from onyx.db.engine import get_async_session
 from onyx.db.engine import get_async_session_with_tenant
-from onyx.db.engine import get_current_tenant_id
 from onyx.db.engine import get_session_with_tenant
 from onyx.db.models import AccessToken
 from onyx.db.models import OAuthAccount
@@ -94,6 +95,7 @@ from onyx.db.models import User
 from onyx.db.users import get_user_by_email
 from onyx.redis.redis_pool import get_async_redis_connection
 from onyx.redis.redis_pool import get_redis_client
+from onyx.server.utils import BasicAuthenticationError
 from onyx.utils.logger import setup_logger
 from onyx.utils.telemetry import create_milestone_and_report
 from onyx.utils.telemetry import optional_telemetry
@@ -103,15 +105,11 @@ from onyx.utils.variable_functionality import fetch_versioned_implementation
 from shared_configs.configs import async_return_default_schema
 from shared_configs.configs import MULTI_TENANT
 from shared_configs.contextvars import CURRENT_TENANT_ID_CONTEXTVAR
+from shared_configs.contextvars import get_current_tenant_id
 
 logger = setup_logger()
 
 
-class BasicAuthenticationError(HTTPException):
-    def __init__(self, detail: str):
-        super().__init__(status_code=status.HTTP_403_FORBIDDEN, detail=detail)
-
-
 def is_user_admin(user: User | None) -> bool:
     if AUTH_TYPE == AuthType.DISABLED:
         return True
@@ -143,6 +141,30 @@ def get_display_email(email: str | None, space_less: bool = False) -> str:
     return email or ""
 
 
+def generate_password() -> str:
+    lowercase_letters = string.ascii_lowercase
+    uppercase_letters = string.ascii_uppercase
+    digits = string.digits
+    special_characters = string.punctuation
+
+    # Ensure at least one of each required character type
+    password = [
+        secrets.choice(uppercase_letters),
+        secrets.choice(digits),
+        secrets.choice(special_characters),
+    ]
+
+    # Fill the rest with a mix of characters
+    remaining_length = 12 - len(password)
+    all_characters = lowercase_letters + uppercase_letters + digits + special_characters
+    password.extend(secrets.choice(all_characters) for _ in range(remaining_length))
+
+    # Shuffle the password to randomize the position of the required characters
+    random.shuffle(password)
+
+    return "".join(password)
+
+
 def user_needs_to_be_verified() -> bool:
     if AUTH_TYPE == AuthType.BASIC or AUTH_TYPE == AuthType.CLOUD:
         return REQUIRE_EMAIL_VERIFICATION
@@ -192,8 +214,8 @@ def verify_email_is_invited(email: str) -> None:
     raise PermissionError("User not on allowed user whitelist")
 
 
-def verify_email_in_whitelist(email: str, tenant_id: str | None = None) -> None:
-    with get_session_with_tenant(tenant_id) as db_session:
+def verify_email_in_whitelist(email: str, tenant_id: str) -> None:
+    with get_session_with_tenant(tenant_id=tenant_id) as db_session:
         if not get_user_by_email(email, db_session):
             verify_email_is_invited(email)
 
@@ -389,7 +411,7 @@ class UserManager(UUIDIDMixin, BaseUserManager[User, uuid.UUID]):
                 "refresh_token": refresh_token,
             }
 
-            user: User
+            user: User | None = None
 
             try:
                 # Attempt to get user by OAuth account
@@ -398,15 +420,20 @@ class UserManager(UUIDIDMixin, BaseUserManager[User, uuid.UUID]):
             except exceptions.UserNotExists:
                 try:
                     # Attempt to get user by email
-                    user = await self.get_by_email(account_email)
+                    user = await self.user_db.get_by_email(account_email)
                     if not associate_by_email:
                         raise exceptions.UserAlreadyExists()
 
-                    user = await self.user_db.add_oauth_account(
-                        user, oauth_account_dict
-                    )
+                    # Make sure user is not None before adding OAuth account
+                    if user is not None:
+                        user = await self.user_db.add_oauth_account(
+                            user, oauth_account_dict
+                        )
+                    else:
+                        # This shouldn't happen since get_by_email would raise UserNotExists
+                        # but adding as a safeguard
+                        raise exceptions.UserNotExists()
 
-                    # If user not found by OAuth account or email, create a new user
                 except exceptions.UserNotExists:
                     password = self.password_helper.generate()
                     user_dict = {
@@ -417,26 +444,36 @@ class UserManager(UUIDIDMixin, BaseUserManager[User, uuid.UUID]):
 
                     user = await self.user_db.create(user_dict)
 
-                    # Explicitly set the Postgres schema for this session to ensure
-                    # OAuth account creation happens in the correct tenant schema
-
-                    # Add OAuth account
-                    await self.user_db.add_oauth_account(user, oauth_account_dict)
-                    await self.on_after_register(user, request)
+                    # Add OAuth account only if user creation was successful
+                    if user is not None:
+                        await self.user_db.add_oauth_account(user, oauth_account_dict)
+                        await self.on_after_register(user, request)
+                    else:
+                        raise HTTPException(
+                            status_code=500, detail="Failed to create user account"
+                        )
 
             else:
-                for existing_oauth_account in user.oauth_accounts:
-                    if (
-                        existing_oauth_account.account_id == account_id
-                        and existing_oauth_account.oauth_name == oauth_name
-                    ):
-                        user = await self.user_db.update_oauth_account(
-                            user,
-                            # NOTE: OAuthAccount DOES implement the OAuthAccountProtocol
-                            # but the type checker doesn't know that :(
-                            existing_oauth_account,  # type: ignore
-                            oauth_account_dict,
-                        )
+                # User exists, update OAuth account if needed
+                if user is not None:  # Add explicit check
+                    for existing_oauth_account in user.oauth_accounts:
+                        if (
+                            existing_oauth_account.account_id == account_id
+                            and existing_oauth_account.oauth_name == oauth_name
+                        ):
+                            user = await self.user_db.update_oauth_account(
+                                user,
+                                # NOTE: OAuthAccount DOES implement the OAuthAccountProtocol
+                                # but the type checker doesn't know that :(
+                                existing_oauth_account,  # type: ignore
+                                oauth_account_dict,
+                            )
+
+            # Ensure user is not None before proceeding
+            if user is None:
+                raise HTTPException(
+                    status_code=500, detail="Failed to authenticate or create user"
+                )
 
             # NOTE: Most IdPs have very short expiry times, and we don't want to force the user to
             # re-authenticate that frequently, so by default this is disabled
@@ -486,6 +523,7 @@ class UserManager(UUIDIDMixin, BaseUserManager[User, uuid.UUID]):
         token = CURRENT_TENANT_ID_CONTEXTVAR.set(tenant_id)
         try:
             user_count = await get_user_count()
+            logger.debug(f"Current tenant user count: {user_count}")
 
             with get_session_with_tenant(tenant_id=tenant_id) as db_session:
                 if user_count == 1:
@@ -507,7 +545,7 @@ class UserManager(UUIDIDMixin, BaseUserManager[User, uuid.UUID]):
         finally:
             CURRENT_TENANT_ID_CONTEXTVAR.reset(token)
 
-        logger.notice(f"User {user.id} has registered.")
+        logger.debug(f"User {user.id} has registered.")
         optional_telemetry(
             record_type=RecordType.SIGN_UP,
             data={"action": "create"},
@@ -531,7 +569,7 @@ class UserManager(UUIDIDMixin, BaseUserManager[User, uuid.UUID]):
             async_return_default_schema,
         )(email=user.email)
 
-        send_forgot_password_email(user.email, token, tenant_id=tenant_id)
+        send_forgot_password_email(user.email, tenant_id=tenant_id, token=token)
 
     async def on_after_request_verify(
         self, user: User, token: str, request: Optional[Request] = None
@@ -549,14 +587,20 @@ class UserManager(UUIDIDMixin, BaseUserManager[User, uuid.UUID]):
     ) -> Optional[User]:
         email = credentials.username
 
-        # Get tenant_id from mapping table
-        tenant_id = await fetch_ee_implementation_or_noop(
-            "onyx.server.tenants.provisioning",
-            "get_or_provision_tenant",
-            async_return_default_schema,
-        )(
-            email=email,
-        )
+        tenant_id: str | None = None
+        try:
+            tenant_id = fetch_ee_implementation_or_noop(
+                "onyx.server.tenants.provisioning",
+                "get_tenant_id_for_email",
+                None,
+            )(
+                email=email,
+            )
+        except Exception as e:
+            logger.warning(
+                f"User attempted to login with invalid credentials: {str(e)}"
+            )
+
         if not tenant_id:
             # User not found in mapping
             self.password_helper.hash(credentials.password)
@@ -595,6 +639,39 @@ class UserManager(UUIDIDMixin, BaseUserManager[User, uuid.UUID]):
 
             return user
 
+    async def reset_password_as_admin(self, user_id: uuid.UUID) -> str:
+        """Admin-only. Generate a random password for a user and return it."""
+        user = await self.get(user_id)
+        new_password = generate_password()
+        await self._update(user, {"password": new_password})
+        return new_password
+
+    async def change_password_if_old_matches(
+        self, user: User, old_password: str, new_password: str
+    ) -> None:
+        """
+        For normal users to change password if they know the old one.
+        Raises 400 if old password doesn't match.
+        """
+        verified, updated_password_hash = self.password_helper.verify_and_update(
+            old_password, user.hashed_password
+        )
+        if not verified:
+            # Raise some HTTPException (or your custom exception) if old password is invalid:
+            from fastapi import HTTPException, status
+
+            raise HTTPException(
+                status_code=status.HTTP_400_BAD_REQUEST,
+                detail="Invalid current password",
+            )
+
+        # If the hash was upgraded behind the scenes, we can keep it before setting the new password:
+        if updated_password_hash:
+            user.hashed_password = updated_password_hash
+
+        # Now apply and validate the new password
+        await self._update(user, {"password": new_password})
+
 
 async def get_user_manager(
     user_db: SQLAlchemyUserDatabase = Depends(get_user_db),
@@ -819,8 +896,9 @@ async def current_limited_user(
 
 async def current_chat_accesssible_user(
     user: User | None = Depends(optional_user),
-    tenant_id: str | None = Depends(get_current_tenant_id),
 ) -> User | None:
+    tenant_id = get_current_tenant_id()
+
     return await double_check_user(
         user, allow_anonymous_access=anonymous_user_enabled(tenant_id=tenant_id)
     )

backend/onyx/background/celery/apps/app_base.py

@@ -2,6 +2,7 @@ import logging
 import multiprocessing
 import time
 from typing import Any
+from typing import cast
 
 import sentry_sdk
 from celery import Task
@@ -33,6 +34,7 @@ from onyx.redis.redis_connector_ext_group_sync import RedisConnectorExternalGrou
 from onyx.redis.redis_connector_prune import RedisConnectorPrune
 from onyx.redis.redis_document_set import RedisDocumentSet
 from onyx.redis.redis_pool import get_redis_client
+from onyx.redis.redis_pool import get_shared_redis_client
 from onyx.redis.redis_usergroup import RedisUserGroup
 from onyx.utils.logger import ColoredFormatter
 from onyx.utils.logger import PlainFormatter
@@ -58,13 +60,35 @@ else:
     logger.debug("Sentry DSN not provided, skipping Sentry initialization")
 
 
+class TenantAwareTask(Task):
+    """A custom base Task that sets tenant_id in a contextvar before running."""
+
+    abstract = True  # So Celery knows not to register this as a real task.
+
+    def __call__(self, *args: Any, **kwargs: Any) -> Any:
+        # Grab tenant_id from the kwargs, or fallback to default if missing.
+        tenant_id = kwargs.get("tenant_id", None) or POSTGRES_DEFAULT_SCHEMA
+
+        # Set the context var
+        CURRENT_TENANT_ID_CONTEXTVAR.set(tenant_id)
+
+        # Actually run the task now
+        try:
+            return super().__call__(*args, **kwargs)
+        finally:
+            # Clear or reset after the task runs
+            # so it does not leak into any subsequent tasks on the same worker process
+            CURRENT_TENANT_ID_CONTEXTVAR.set(None)
+
+
+@task_prerun.connect
 def on_task_prerun(
     sender: Any | None = None,
     task_id: str | None = None,
     task: Task | None = None,
     args: tuple[Any, ...] | None = None,
     kwargs: dict[str, Any] | None = None,
-    **kwds: Any,
+    **other_kwargs: Any,
 ) -> None:
     pass
 
@@ -108,9 +132,9 @@ def on_task_postrun(
     # Get tenant_id directly from kwargs- each celery task has a tenant_id kwarg
     if not kwargs:
         logger.error(f"Task {task.name} (ID: {task_id}) is missing kwargs")
-        tenant_id = None
+        tenant_id = POSTGRES_DEFAULT_SCHEMA
     else:
-        tenant_id = kwargs.get("tenant_id")
+        tenant_id = cast(str, kwargs.get("tenant_id", POSTGRES_DEFAULT_SCHEMA))
 
     task_logger.debug(
         f"Task {task.name} (ID: {task_id}) completed with state: {state} "
@@ -201,7 +225,7 @@ def wait_for_redis(sender: Any, **kwargs: Any) -> None:
     Will raise WorkerShutdown to kill the celery worker if the timeout
     is reached."""
 
-    r = get_redis_client(tenant_id=None)
+    r = get_shared_redis_client()
 
     WAIT_INTERVAL = 5
     WAIT_LIMIT = 60
@@ -287,7 +311,7 @@ def on_secondary_worker_init(sender: Any, **kwargs: Any) -> None:
     # Set up variables for waiting on primary worker
     WAIT_INTERVAL = 5
     WAIT_LIMIT = 60
-    r = get_redis_client(tenant_id=None)
+    r = get_shared_redis_client()
     time_start = time.monotonic()
 
     logger.info("Waiting for primary worker to be ready...")
@@ -439,24 +463,6 @@ class TenantContextFilter(logging.Filter):
         return True
 
 
-@task_prerun.connect
-def set_tenant_id(
-    sender: Any | None = None,
-    task_id: str | None = None,
-    task: Task | None = None,
-    args: tuple[Any, ...] | None = None,
-    kwargs: dict[str, Any] | None = None,
-    **other_kwargs: Any,
-) -> None:
-    """Signal handler to set tenant ID in context var before task starts."""
-    tenant_id = (
-        kwargs.get("tenant_id", POSTGRES_DEFAULT_SCHEMA)
-        if kwargs
-        else POSTGRES_DEFAULT_SCHEMA
-    )
-    CURRENT_TENANT_ID_CONTEXTVAR.set(tenant_id)
-
-
 @task_postrun.connect
 def reset_tenant_id(
     sender: Any | None = None,

backend/onyx/background/celery/apps/beat.py

@@ -132,6 +132,7 @@ class DynamicTenantScheduler(PersistentScheduler):
                         f"Adding options to task {tenant_task_name}: {options}"
                     )
                     tenant_task["options"] = options
+
                 new_schedule[tenant_task_name] = tenant_task
 
         return new_schedule
@@ -256,3 +257,4 @@ def on_setup_logging(
 
 
 celery_app.conf.beat_scheduler = DynamicTenantScheduler
+celery_app.conf.task_default_base = app_base.TenantAwareTask

backend/onyx/background/celery/apps/heavy.py

@@ -20,6 +20,7 @@ logger = setup_logger()
 
 celery_app = Celery(__name__)
 celery_app.config_from_object("onyx.background.celery.configs.heavy")
+celery_app.Task = app_base.TenantAwareTask  # type: ignore [misc]
 
 
 @signals.task_prerun.connect

backend/onyx/background/celery/apps/indexing.py

@@ -21,6 +21,7 @@ logger = setup_logger()
 
 celery_app = Celery(__name__)
 celery_app.config_from_object("onyx.background.celery.configs.indexing")
+celery_app.Task = app_base.TenantAwareTask  # type: ignore [misc]
 
 
 @signals.task_prerun.connect

backend/onyx/background/celery/apps/light.py

@@ -23,6 +23,7 @@ logger = setup_logger()
 
 celery_app = Celery(__name__)
 celery_app.config_from_object("onyx.background.celery.configs.light")
+celery_app.Task = app_base.TenantAwareTask  # type: ignore [misc]
 
 
 @signals.task_prerun.connect

backend/onyx/background/celery/apps/monitoring.py

@@ -20,6 +20,7 @@ logger = setup_logger()
 
 celery_app = Celery(__name__)
 celery_app.config_from_object("onyx.background.celery.configs.monitoring")
+celery_app.Task = app_base.TenantAwareTask  # type: ignore [misc]
 
 
 @signals.task_prerun.connect

backend/onyx/background/celery/apps/primary.py

@@ -24,7 +24,7 @@ from onyx.configs.constants import CELERY_PRIMARY_WORKER_LOCK_TIMEOUT
 from onyx.configs.constants import OnyxRedisConstants
 from onyx.configs.constants import OnyxRedisLocks
 from onyx.configs.constants import POSTGRES_CELERY_WORKER_PRIMARY_APP_NAME
-from onyx.db.engine import get_session_with_default_tenant
+from onyx.db.engine import get_session_with_current_tenant
 from onyx.db.engine import SqlEngine
 from onyx.db.index_attempt import get_index_attempt
 from onyx.db.index_attempt import mark_attempt_canceled
@@ -38,7 +38,7 @@ from onyx.redis.redis_connector_index import RedisConnectorIndex
 from onyx.redis.redis_connector_prune import RedisConnectorPrune
 from onyx.redis.redis_connector_stop import RedisConnectorStop
 from onyx.redis.redis_document_set import RedisDocumentSet
-from onyx.redis.redis_pool import get_redis_client
+from onyx.redis.redis_pool import get_shared_redis_client
 from onyx.redis.redis_usergroup import RedisUserGroup
 from onyx.utils.logger import setup_logger
 from shared_configs.configs import MULTI_TENANT
@@ -47,6 +47,7 @@ logger = setup_logger()
 
 celery_app = Celery(__name__)
 celery_app.config_from_object("onyx.background.celery.configs.primary")
+celery_app.Task = app_base.TenantAwareTask  # type: ignore [misc]
 
 
 @signals.task_prerun.connect
@@ -101,7 +102,7 @@ def on_worker_init(sender: Worker, **kwargs: Any) -> None:
 
     # This is singleton work that should be done on startup exactly once
     # by the primary worker. This is unnecessary in the multi tenant scenario
-    r = get_redis_client(tenant_id=None)
+    r = get_shared_redis_client()
 
     # Log the role and slave count - being connected to a slave or slave count > 0 could be problematic
     info: dict[str, Any] = cast(dict, r.info("replication"))
@@ -158,7 +159,7 @@ def on_worker_init(sender: Worker, **kwargs: Any) -> None:
     RedisConnectorExternalGroupSync.reset_all(r)
 
     # mark orphaned index attempts as failed
-    with get_session_with_default_tenant() as db_session:
+    with get_session_with_current_tenant() as db_session:
         unfenced_attempt_ids = get_unfenced_index_attempt_ids(db_session, r)
         for attempt_id in unfenced_attempt_ids:
             attempt = get_index_attempt(db_session, attempt_id)
@@ -234,7 +235,7 @@ class HubPeriodicTask(bootsteps.StartStopStep):
 
             lock: RedisLock = worker.primary_worker_lock
 
-            r = get_redis_client(tenant_id=None)
+            r = get_shared_redis_client()
 
             if lock.owned():
                 task_logger.debug("Reacquiring primary worker lock.")

backend/onyx/background/celery/celery_redis.py

@@ -92,7 +92,8 @@ def celery_find_task(task_id: str, queue: str, r: Redis) -> int:
 
 
 def celery_get_queued_task_ids(queue: str, r: Redis) -> set[str]:
-    """This is a redis specific way to build a list of tasks in a queue.
+    """This is a redis specific way to build a list of tasks in a queue and return them
+    as a set.
 
     This helps us read the queue once and then efficiently look for missing tasks
     in the queue.

backend/onyx/background/celery/celery_utils.py

@@ -34,7 +34,7 @@ def _get_deletion_status(
     connector_id: int,
     credential_id: int,
     db_session: Session,
-    tenant_id: str | None = None,
+    tenant_id: str,
 ) -> TaskQueueState | None:
     """We no longer store TaskQueueState in the DB for a deletion attempt.
     This function populates TaskQueueState by just checking redis.
@@ -67,7 +67,7 @@ def get_deletion_attempt_snapshot(
     connector_id: int,
     credential_id: int,
     db_session: Session,
-    tenant_id: str | None = None,
+    tenant_id: str,
 ) -> DeletionAttemptSnapshot | None:
     deletion_task = _get_deletion_status(
         connector_id, credential_id, db_session, tenant_id

backend/onyx/background/celery/tasks/beat_schedule.py

@@ -36,6 +36,15 @@ beat_task_templates.extend(
                 "expires": BEAT_EXPIRES_DEFAULT,
             },
         },
+        {
+            "name": "check-for-checkpoint-cleanup",
+            "task": OnyxCeleryTask.CHECK_FOR_CHECKPOINT_CLEANUP,
+            "schedule": timedelta(hours=1),
+            "options": {
+                "priority": OnyxCeleryPriority.LOW,
+                "expires": BEAT_EXPIRES_DEFAULT,
+            },
+        },
         {
             "name": "check-for-connector-deletion",
             "task": OnyxCeleryTask.CHECK_FOR_CONNECTOR_DELETION,

backend/onyx/background/celery/tasks/connector_deletion/tasks.py

@@ -8,16 +8,21 @@ from celery import Celery
 from celery import shared_task
 from celery import Task
 from celery.exceptions import SoftTimeLimitExceeded
+from pydantic import ValidationError
 from redis import Redis
 from redis.lock import Lock as RedisLock
 from sqlalchemy.orm import Session
 
 from onyx.background.celery.apps.app_base import task_logger
+from onyx.background.celery.celery_redis import celery_get_queue_length
+from onyx.background.celery.celery_redis import celery_get_queued_task_ids
 from onyx.configs.app_configs import JOB_TIMEOUT
 from onyx.configs.constants import CELERY_GENERIC_BEAT_LOCK_TIMEOUT
+from onyx.configs.constants import OnyxCeleryQueues
 from onyx.configs.constants import OnyxCeleryTask
 from onyx.configs.constants import OnyxRedisConstants
 from onyx.configs.constants import OnyxRedisLocks
+from onyx.configs.constants import OnyxRedisSignals
 from onyx.db.connector import fetch_connector_by_id
 from onyx.db.connector_credential_pair import add_deletion_failure_message
 from onyx.db.connector_credential_pair import (
@@ -27,7 +32,7 @@ from onyx.db.connector_credential_pair import get_connector_credential_pair_from
 from onyx.db.connector_credential_pair import get_connector_credential_pairs
 from onyx.db.document import get_document_ids_for_connector_credential_pair
 from onyx.db.document_set import delete_document_set_cc_pair_relationship__no_commit
-from onyx.db.engine import get_session_with_tenant
+from onyx.db.engine import get_session_with_current_tenant
 from onyx.db.enums import ConnectorCredentialPairStatus
 from onyx.db.enums import SyncStatus
 from onyx.db.enums import SyncType
@@ -52,6 +57,51 @@ class TaskDependencyError(RuntimeError):
     with connector deletion."""
 
 
+def revoke_tasks_blocking_deletion(
+    redis_connector: RedisConnector, db_session: Session, app: Celery
+) -> None:
+    search_settings_list = get_all_search_settings(db_session)
+    for search_settings in search_settings_list:
+        redis_connector_index = redis_connector.new_index(search_settings.id)
+        try:
+            index_payload = redis_connector_index.payload
+            if index_payload and index_payload.celery_task_id:
+                app.control.revoke(index_payload.celery_task_id)
+                task_logger.info(
+                    f"Revoked indexing task {index_payload.celery_task_id}."
+                )
+        except Exception:
+            task_logger.exception("Exception while revoking indexing task")
+
+    try:
+        permissions_sync_payload = redis_connector.permissions.payload
+        if permissions_sync_payload and permissions_sync_payload.celery_task_id:
+            app.control.revoke(permissions_sync_payload.celery_task_id)
+            task_logger.info(
+                f"Revoked permissions sync task {permissions_sync_payload.celery_task_id}."
+            )
+    except Exception:
+        task_logger.exception("Exception while revoking pruning task")
+
+    try:
+        prune_payload = redis_connector.prune.payload
+        if prune_payload and prune_payload.celery_task_id:
+            app.control.revoke(prune_payload.celery_task_id)
+            task_logger.info(f"Revoked pruning task {prune_payload.celery_task_id}.")
+    except Exception:
+        task_logger.exception("Exception while revoking permissions sync task")
+
+    try:
+        external_group_sync_payload = redis_connector.external_group_sync.payload
+        if external_group_sync_payload and external_group_sync_payload.celery_task_id:
+            app.control.revoke(external_group_sync_payload.celery_task_id)
+            task_logger.info(
+                f"Revoked external group sync task {external_group_sync_payload.celery_task_id}."
+            )
+    except Exception:
+        task_logger.exception("Exception while revoking external group sync task")
+
+
 @shared_task(
     name=OnyxCeleryTask.CHECK_FOR_CONNECTOR_DELETION,
     ignore_result=True,
@@ -59,32 +109,46 @@ class TaskDependencyError(RuntimeError):
     trail=False,
     bind=True,
 )
-def check_for_connector_deletion_task(
-    self: Task, *, tenant_id: str | None
-) -> bool | None:
-    r = get_redis_client(tenant_id=tenant_id)
-    r_replica = get_redis_replica_client(tenant_id=tenant_id)
+def check_for_connector_deletion_task(self: Task, *, tenant_id: str) -> bool | None:
+    r = get_redis_client()
+    r_replica = get_redis_replica_client()
+    r_celery: Redis = self.app.broker_connection().channel().client  # type: ignore
 
     lock_beat: RedisLock = r.lock(
         OnyxRedisLocks.CHECK_CONNECTOR_DELETION_BEAT_LOCK,
         timeout=CELERY_GENERIC_BEAT_LOCK_TIMEOUT,
     )
 
-    # these tasks should never overlap
+    # Prevent this task from overlapping with itself
     if not lock_beat.acquire(blocking=False):
         return None
 
     try:
+        # we want to run this less frequently than the overall task
+        lock_beat.reacquire()
+        if not r.exists(OnyxRedisSignals.BLOCK_VALIDATE_CONNECTOR_DELETION_FENCES):
+            # clear fences that don't have associated celery tasks in progress
+            try:
+                validate_connector_deletion_fences(
+                    tenant_id, r, r_replica, r_celery, lock_beat
+                )
+            except Exception:
+                task_logger.exception(
+                    "Exception while validating connector deletion fences"
+                )
+
+            r.set(OnyxRedisSignals.BLOCK_VALIDATE_CONNECTOR_DELETION_FENCES, 1, ex=300)
+
         # collect cc_pair_ids
         cc_pair_ids: list[int] = []
-        with get_session_with_tenant(tenant_id) as db_session:
+        with get_session_with_current_tenant() as db_session:
             cc_pairs = get_connector_credential_pairs(db_session)
             for cc_pair in cc_pairs:
                 cc_pair_ids.append(cc_pair.id)
 
         # try running cleanup on the cc_pair_ids
         for cc_pair_id in cc_pair_ids:
-            with get_session_with_tenant(tenant_id) as db_session:
+            with get_session_with_current_tenant() as db_session:
                 redis_connector = RedisConnector(tenant_id, cc_pair_id)
                 try:
                     try_generate_document_cc_pair_cleanup_tasks(
@@ -92,9 +156,38 @@ def check_for_connector_deletion_task(
                     )
                 except TaskDependencyError as e:
                     # this means we wanted to start deleting but dependent tasks were running
-                    # Leave a stop signal to clear indexing and pruning tasks more quickly
+                    # on the first error, we set a stop signal and revoke the dependent tasks
+                    # on subsequent errors, we hard reset blocking fences after our specified timeout
+                    # is exceeded
                     task_logger.info(str(e))
-                    redis_connector.stop.set_fence(True)
+
+                    if not redis_connector.stop.fenced:
+                        # one time revoke of celery tasks
+                        task_logger.info("Revoking any tasks blocking deletion.")
+                        revoke_tasks_blocking_deletion(
+                            redis_connector, db_session, self.app
+                        )
+                        redis_connector.stop.set_fence(True)
+                        redis_connector.stop.set_timeout()
+                    else:
+                        # stop signal already set
+                        if redis_connector.stop.timed_out:
+                            # waiting too long, just reset blocking fences
+                            task_logger.info(
+                                "Timed out waiting for tasks blocking deletion. Resetting blocking fences."
+                            )
+                            search_settings_list = get_all_search_settings(db_session)
+                            for search_settings in search_settings_list:
+                                redis_connector_index = redis_connector.new_index(
+                                    search_settings.id
+                                )
+                                redis_connector_index.reset()
+                            redis_connector.prune.reset()
+                            redis_connector.permissions.reset()
+                            redis_connector.external_group_sync.reset()
+                        else:
+                            # just wait
+                            pass
                 else:
                     # clear the stop signal if it exists ... no longer needed
                     redis_connector.stop.set_fence(False)
@@ -129,7 +222,7 @@ def try_generate_document_cc_pair_cleanup_tasks(
     cc_pair_id: int,
     db_session: Session,
     lock_beat: RedisLock,
-    tenant_id: str | None,
+    tenant_id: str,
 ) -> int | None:
     """Returns an int if syncing is needed. The int represents the number of sync tasks generated.
     Note that syncing can still be required even if the number of sync tasks generated is zero.
@@ -169,6 +262,7 @@ def try_generate_document_cc_pair_cleanup_tasks(
         return None
 
     # set a basic fence to start
+    redis_connector.delete.set_active()
     fence_payload = RedisConnectorDeletePayload(
         num_tasks=None,
         submitted=datetime.now(timezone.utc),
@@ -249,7 +343,7 @@ def try_generate_document_cc_pair_cleanup_tasks(
 
 
 def monitor_connector_deletion_taskset(
-    tenant_id: str | None, key_bytes: bytes, r: Redis
+    tenant_id: str, key_bytes: bytes, r: Redis
 ) -> None:
     fence_key = key_bytes.decode("utf-8")
     cc_pair_id_str = RedisConnector.get_id_from_fence_key(fence_key)
@@ -277,7 +371,7 @@ def monitor_connector_deletion_taskset(
         f"Connector deletion progress: cc_pair={cc_pair_id} remaining={remaining} initial={fence_data.num_tasks}"
     )
     if remaining > 0:
-        with get_session_with_tenant(tenant_id) as db_session:
+        with get_session_with_current_tenant() as db_session:
             update_sync_record_status(
                 db_session=db_session,
                 entity_id=cc_pair_id,
@@ -287,7 +381,7 @@ def monitor_connector_deletion_taskset(
             )
         return
 
-    with get_session_with_tenant(tenant_id) as db_session:
+    with get_session_with_current_tenant() as db_session:
         cc_pair = get_connector_credential_pair_from_id(
             db_session=db_session,
             cc_pair_id=cc_pair_id,
@@ -401,3 +495,171 @@ def monitor_connector_deletion_taskset(
     )
 
     redis_connector.delete.reset()
+
+
+def validate_connector_deletion_fences(
+    tenant_id: str,
+    r: Redis,
+    r_replica: Redis,
+    r_celery: Redis,
+    lock_beat: RedisLock,
+) -> None:
+    # building lookup table can be expensive, so we won't bother
+    # validating until the queue is small
+    CONNECTION_DELETION_VALIDATION_MAX_QUEUE_LEN = 1024
+
+    queue_len = celery_get_queue_length(OnyxCeleryQueues.CONNECTOR_DELETION, r_celery)
+    if queue_len > CONNECTION_DELETION_VALIDATION_MAX_QUEUE_LEN:
+        return
+
+    queued_upsert_tasks = celery_get_queued_task_ids(
+        OnyxCeleryQueues.CONNECTOR_DELETION, r_celery
+    )
+
+    # validate all existing connector deletion jobs
+    lock_beat.reacquire()
+    keys = cast(set[Any], r_replica.smembers(OnyxRedisConstants.ACTIVE_FENCES))
+    for key in keys:
+        key_bytes = cast(bytes, key)
+        key_str = key_bytes.decode("utf-8")
+        if not key_str.startswith(RedisConnectorDelete.FENCE_PREFIX):
+            continue
+
+        validate_connector_deletion_fence(
+            tenant_id,
+            key_bytes,
+            queued_upsert_tasks,
+            r,
+        )
+
+        lock_beat.reacquire()
+
+    return
+
+
+def validate_connector_deletion_fence(
+    tenant_id: str,
+    key_bytes: bytes,
+    queued_tasks: set[str],
+    r: Redis,
+) -> None:
+    """Checks for the error condition where an indexing fence is set but the associated celery tasks don't exist.
+    This can happen if the indexing worker hard crashes or is terminated.
+    Being in this bad state means the fence will never clear without help, so this function
+    gives the help.
+
+    How this works:
+    1. This function renews the active signal with a 5 minute TTL under the following conditions
+    1.2. When the task is seen in the redis queue
+    1.3. When the task is seen in the reserved / prefetched list
+
+    2. Externally, the active signal is renewed when:
+    2.1. The fence is created
+    2.2. The indexing watchdog checks the spawned task.
+
+    3. The TTL allows us to get through the transitions on fence startup
+    and when the task starts executing.
+
+    More TTL clarification: it is seemingly impossible to exactly query Celery for
+    whether a task is in the queue or currently executing.
+    1. An unknown task id is always returned as state PENDING.
+    2. Redis can be inspected for the task id, but the task id is gone between the time a worker receives the task
+    and the time it actually starts on the worker.
+
+    queued_tasks: the celery queue of lightweight permission sync tasks
+    reserved_tasks: prefetched tasks for sync task generator
+    """
+    # if the fence doesn't exist, there's nothing to do
+    fence_key = key_bytes.decode("utf-8")
+    cc_pair_id_str = RedisConnector.get_id_from_fence_key(fence_key)
+    if cc_pair_id_str is None:
+        task_logger.warning(
+            f"validate_connector_deletion_fence - could not parse id from {fence_key}"
+        )
+        return
+
+    cc_pair_id = int(cc_pair_id_str)
+    # parse out metadata and initialize the helper class with it
+    redis_connector = RedisConnector(tenant_id, int(cc_pair_id))
+
+    # check to see if the fence/payload exists
+    if not redis_connector.delete.fenced:
+        return
+
+    # in the cloud, the payload format may have changed ...
+    # it's a little sloppy, but just reset the fence for now if that happens
+    # TODO: add intentional cleanup/abort logic
+    try:
+        payload = redis_connector.delete.payload
+    except ValidationError:
+        task_logger.exception(
+            "validate_connector_deletion_fence - "
+            "Resetting fence because fence schema is out of date: "
+            f"cc_pair={cc_pair_id} "
+            f"fence={fence_key}"
+        )
+
+        redis_connector.delete.reset()
+        return
+
+    if not payload:
+        return
+
+    # OK, there's actually something for us to validate
+
+    # look up every task in the current taskset in the celery queue
+    # every entry in the taskset should have an associated entry in the celery task queue
+    # because we get the celery tasks first, the entries in our own permissions taskset
+    # should be roughly a subset of the tasks in celery
+
+    # this check isn't very exact, but should be sufficient over a period of time
+    # A single successful check over some number of attempts is sufficient.
+
+    # TODO: if the number of tasks in celery is much lower than than the taskset length
+    # we might be able to shortcut the lookup since by definition some of the tasks
+    # must not exist in celery.
+
+    tasks_scanned = 0
+    tasks_not_in_celery = 0  # a non-zero number after completing our check is bad
+
+    for member in r.sscan_iter(redis_connector.delete.taskset_key):
+        tasks_scanned += 1
+
+        member_bytes = cast(bytes, member)
+        member_str = member_bytes.decode("utf-8")
+        if member_str in queued_tasks:
+            continue
+
+        tasks_not_in_celery += 1
+
+    task_logger.info(
+        "validate_connector_deletion_fence task check: "
+        f"tasks_scanned={tasks_scanned} tasks_not_in_celery={tasks_not_in_celery}"
+    )
+
+    # we're active if there are still tasks to run and those tasks all exist in celery
+    if tasks_scanned > 0 and tasks_not_in_celery == 0:
+        redis_connector.delete.set_active()
+        return
+
+    # we may want to enable this check if using the active task list somehow isn't good enough
+    # if redis_connector_index.generator_locked():
+    #     logger.info(f"{payload.celery_task_id} is currently executing.")
+
+    # if we get here, we didn't find any direct indication that the associated celery tasks exist,
+    # but they still might be there due to gaps in our ability to check states during transitions
+    # Checking the active signal safeguards us against these transition periods
+    # (which has a duration that allows us to bridge those gaps)
+    if redis_connector.delete.active():
+        return
+
+    # celery tasks don't exist and the active signal has expired, possibly due to a crash. Clean it up.
+    task_logger.warning(
+        "validate_connector_deletion_fence - "
+        "Resetting fence because no associated celery tasks were found: "
+        f"cc_pair={cc_pair_id} "
+        f"fence={fence_key}"
+    )
+
+    redis_connector.delete.reset()
+    return

backend/onyx/background/celery/tasks/doc_permission_syncing/tasks.py

@@ -30,6 +30,7 @@ from onyx.background.celery.celery_redis import celery_find_task
 from onyx.background.celery.celery_redis import celery_get_queue_length
 from onyx.background.celery.celery_redis import celery_get_queued_task_ids
 from onyx.background.celery.celery_redis import celery_get_unacked_task_ids
+from onyx.background.celery.tasks.shared.tasks import OnyxCeleryTaskCompletionStatus
 from onyx.configs.app_configs import JOB_TIMEOUT
 from onyx.configs.constants import CELERY_GENERIC_BEAT_LOCK_TIMEOUT
 from onyx.configs.constants import CELERY_PERMISSIONS_SYNC_LOCK_TIMEOUT
@@ -42,10 +43,12 @@ from onyx.configs.constants import OnyxCeleryTask
 from onyx.configs.constants import OnyxRedisConstants
 from onyx.configs.constants import OnyxRedisLocks
 from onyx.configs.constants import OnyxRedisSignals
+from onyx.connectors.factory import validate_ccpair_for_user
 from onyx.db.connector import mark_cc_pair_as_permissions_synced
 from onyx.db.connector_credential_pair import get_connector_credential_pair_from_id
+from onyx.db.connector_credential_pair import update_connector_credential_pair
 from onyx.db.document import upsert_document_by_connector_credential_pair
-from onyx.db.engine import get_session_with_tenant
+from onyx.db.engine import get_session_with_current_tenant
 from onyx.db.enums import AccessType
 from onyx.db.enums import ConnectorCredentialPairStatus
 from onyx.db.enums import SyncStatus
@@ -63,6 +66,7 @@ from onyx.redis.redis_pool import get_redis_replica_client
 from onyx.redis.redis_pool import redis_lock_dump
 from onyx.server.utils import make_short_id
 from onyx.utils.logger import doc_permission_sync_ctx
+from onyx.utils.logger import format_error_for_logging
 from onyx.utils.logger import LoggerContextVars
 from onyx.utils.logger import setup_logger
 
@@ -119,13 +123,13 @@ def _is_external_doc_permissions_sync_due(cc_pair: ConnectorCredentialPair) -> b
     soft_time_limit=JOB_TIMEOUT,
     bind=True,
 )
-def check_for_doc_permissions_sync(self: Task, *, tenant_id: str | None) -> bool | None:
+def check_for_doc_permissions_sync(self: Task, *, tenant_id: str) -> bool | None:
     # TODO(rkuo): merge into check function after lookup table for fences is added
 
     # we need to use celery's redis client to access its redis data
     # (which lives on a different db number)
-    r = get_redis_client(tenant_id=tenant_id)
-    r_replica = get_redis_replica_client(tenant_id=tenant_id)
+    r = get_redis_client()
+    r_replica = get_redis_replica_client()
     r_celery: Redis = self.app.broker_connection().channel().client  # type: ignore
 
     lock_beat: RedisLock = r.lock(
@@ -140,7 +144,7 @@ def check_for_doc_permissions_sync(self: Task, *, tenant_id: str | None) -> bool
     try:
         # get all cc pairs that need to be synced
         cc_pair_ids_to_sync: list[int] = []
-        with get_session_with_tenant(tenant_id) as db_session:
+        with get_session_with_current_tenant() as db_session:
             cc_pairs = get_all_auto_sync_cc_pairs(db_session)
 
             for cc_pair in cc_pairs:
@@ -189,16 +193,23 @@ def check_for_doc_permissions_sync(self: Task, *, tenant_id: str | None) -> bool
 
             key_str = key_bytes.decode("utf-8")
             if key_str.startswith(RedisConnectorPermissionSync.FENCE_PREFIX):
-                with get_session_with_tenant(tenant_id) as db_session:
+                with get_session_with_current_tenant() as db_session:
                     monitor_ccpair_permissions_taskset(
                         tenant_id, key_bytes, r, db_session
                     )
+        task_logger.info(f"check_for_doc_permissions_sync finished: tenant={tenant_id}")
     except SoftTimeLimitExceeded:
         task_logger.info(
             "Soft time limit exceeded, task is being terminated gracefully."
         )
-    except Exception:
-        task_logger.exception(f"Unexpected exception: tenant={tenant_id}")
+    except Exception as e:
+        error_msg = format_error_for_logging(e)
+        task_logger.warning(
+            f"Unexpected check_for_doc_permissions_sync exception: tenant={tenant_id} {error_msg}"
+        )
+        task_logger.exception(
+            f"Unexpected check_for_doc_permissions_sync exception: tenant={tenant_id}"
+        )
     finally:
         if lock_beat.owned():
             lock_beat.release()
@@ -210,7 +221,7 @@ def try_creating_permissions_sync_task(
     app: Celery,
     cc_pair_id: int,
     r: Redis,
-    tenant_id: str | None,
+    tenant_id: str,
 ) -> str | None:
     """Returns a randomized payload id on success.
     Returns None if no syncing is required."""
@@ -247,7 +258,7 @@ def try_creating_permissions_sync_task(
         # create before setting fence to avoid race condition where the monitoring
         # task updates the sync record before it is created
         try:
-            with get_session_with_tenant(tenant_id) as db_session:
+            with get_session_with_current_tenant() as db_session:
                 insert_sync_record(
                     db_session=db_session,
                     entity_id=cc_pair_id,
@@ -282,13 +293,19 @@ def try_creating_permissions_sync_task(
         redis_connector.permissions.set_fence(payload)
 
         payload_id = payload.id
-    except Exception:
-        task_logger.exception(f"Unexpected exception: cc_pair={cc_pair_id}")
+    except Exception as e:
+        error_msg = format_error_for_logging(e)
+        task_logger.warning(
+            f"Unexpected try_creating_permissions_sync_task exception: cc_pair={cc_pair_id} {error_msg}"
+        )
         return None
     finally:
         if lock.owned():
             lock.release()
 
+    task_logger.info(
+        f"try_creating_permissions_sync_task finished: cc_pair={cc_pair_id} payload_id={payload_id}"
+    )
     return payload_id
 
 
@@ -303,7 +320,7 @@ def try_creating_permissions_sync_task(
 def connector_permission_sync_generator_task(
     self: Task,
     cc_pair_id: int,
-    tenant_id: str | None,
+    tenant_id: str,
 ) -> None:
     """
     Permission sync task that handles document permission syncing for a given connector credential pair
@@ -321,7 +338,7 @@ def connector_permission_sync_generator_task(
 
     redis_connector = RedisConnector(tenant_id, cc_pair_id)
 
-    r = get_redis_client(tenant_id=tenant_id)
+    r = get_redis_client()
 
     # this wait is needed to avoid a race condition where
     # the primary worker sends the task and it is immediately executed
@@ -378,7 +395,7 @@ def connector_permission_sync_generator_task(
         return None
 
     try:
-        with get_session_with_tenant(tenant_id) as db_session:
+        with get_session_with_current_tenant() as db_session:
             cc_pair = get_connector_credential_pair_from_id(
                 db_session=db_session,
                 cc_pair_id=cc_pair_id,
@@ -388,6 +405,29 @@ def connector_permission_sync_generator_task(
                     f"No connector credential pair found for id: {cc_pair_id}"
                 )
 
+            try:
+                created = validate_ccpair_for_user(
+                    cc_pair.connector.id,
+                    cc_pair.credential.id,
+                    db_session,
+                    enforce_creation=False,
+                )
+                if not created:
+                    task_logger.warning(
+                        f"Unable to create connector credential pair for id: {cc_pair_id}"
+                    )
+            except Exception:
+                task_logger.exception(
+                    f"validate_ccpair_permissions_sync exceptioned: cc_pair={cc_pair_id}"
+                )
+                update_connector_credential_pair(
+                    db_session=db_session,
+                    connector_id=cc_pair.connector.id,
+                    credential_id=cc_pair.credential.id,
+                    status=ConnectorCredentialPairStatus.INVALID,
+                )
+                raise
+
             source_type = cc_pair.connector.source
 
             doc_sync_func = DOC_PERMISSIONS_FUNC_MAP.get(source_type)
@@ -439,6 +479,10 @@ def connector_permission_sync_generator_task(
             redis_connector.permissions.generator_complete = tasks_generated
 
     except Exception as e:
+        error_msg = format_error_for_logging(e)
+        task_logger.warning(
+            f"Permission sync exceptioned: cc_pair={cc_pair_id} payload_id={payload_id} {error_msg}"
+        )
         task_logger.exception(
             f"Permission sync exceptioned: cc_pair={cc_pair_id} payload_id={payload_id}"
         )
@@ -465,7 +509,7 @@ def connector_permission_sync_generator_task(
 )
 def update_external_document_permissions_task(
     self: Task,
-    tenant_id: str | None,
+    tenant_id: str,
     serialized_doc_external_access: dict,
     source_string: str,
     connector_id: int,
@@ -473,6 +517,8 @@ def update_external_document_permissions_task(
 ) -> bool:
     start = time.monotonic()
 
+    completion_status = OnyxCeleryTaskCompletionStatus.UNDEFINED
+
     document_external_access = DocExternalAccess.from_dict(
         serialized_doc_external_access
     )
@@ -480,7 +526,8 @@ def update_external_document_permissions_task(
     external_access = document_external_access.external_access
 
     try:
-        with get_session_with_tenant(tenant_id) as db_session:
+        with get_session_with_current_tenant() as db_session:
+            # Add the users to the DB if they don't exist
             batch_add_ext_perm_user_if_not_exists(
                 db_session=db_session,
                 emails=list(external_access.external_user_emails),
@@ -511,18 +558,33 @@ def update_external_document_permissions_task(
                 f"elapsed={elapsed:.2f}"
             )
 
-    except Exception:
+        completion_status = OnyxCeleryTaskCompletionStatus.SUCCEEDED
+    except Exception as e:
+        error_msg = format_error_for_logging(e)
+        task_logger.warning(
+            f"Exception in update_external_document_permissions_task: connector_id={connector_id} doc_id={doc_id} {error_msg}"
+        )
         task_logger.exception(
-            f"Exception in update_external_document_permissions_task: "
+            f"update_external_document_permissions_task exceptioned: "
             f"connector_id={connector_id} doc_id={doc_id}"
         )
+        completion_status = OnyxCeleryTaskCompletionStatus.NON_RETRYABLE_EXCEPTION
+    finally:
+        task_logger.info(
+            f"update_external_document_permissions_task completed: status={completion_status.value} doc={doc_id}"
+        )
+
+    if completion_status != OnyxCeleryTaskCompletionStatus.SUCCEEDED:
         return False
 
+    task_logger.info(
+        f"update_external_document_permissions_task finished: connector_id={connector_id} doc_id={doc_id}"
+    )
     return True
 
 
 def validate_permission_sync_fences(
-    tenant_id: str | None,
+    tenant_id: str,
     r: Redis,
     r_replica: Redis,
     r_celery: Redis,
@@ -569,7 +631,7 @@ def validate_permission_sync_fences(
 
 
 def validate_permission_sync_fence(
-    tenant_id: str | None,
+    tenant_id: str,
     key_bytes: bytes,
     queued_tasks: set[str],
     reserved_tasks: set[str],
@@ -779,7 +841,7 @@ class PermissionSyncCallback(IndexingHeartbeatInterface):
 
 
 def monitor_ccpair_permissions_taskset(
-    tenant_id: str | None, key_bytes: bytes, r: Redis, db_session: Session
+    tenant_id: str, key_bytes: bytes, r: Redis, db_session: Session
 ) -> None:
     fence_key = key_bytes.decode("utf-8")
     cc_pair_id_str = RedisConnector.get_id_from_fence_key(fence_key)

backend/onyx/background/celery/tasks/external_group_syncing/tasks.py

@@ -37,9 +37,12 @@ from onyx.configs.constants import OnyxCeleryTask
 from onyx.configs.constants import OnyxRedisConstants
 from onyx.configs.constants import OnyxRedisLocks
 from onyx.configs.constants import OnyxRedisSignals
+from onyx.connectors.exceptions import ConnectorValidationError
+from onyx.connectors.factory import validate_ccpair_for_user
 from onyx.db.connector import mark_cc_pair_as_external_group_synced
 from onyx.db.connector_credential_pair import get_connector_credential_pair_from_id
-from onyx.db.engine import get_session_with_tenant
+from onyx.db.connector_credential_pair import update_connector_credential_pair
+from onyx.db.engine import get_session_with_current_tenant
 from onyx.db.enums import AccessType
 from onyx.db.enums import ConnectorCredentialPairStatus
 from onyx.db.enums import SyncStatus
@@ -55,6 +58,7 @@ from onyx.redis.redis_connector_ext_group_sync import (
 from onyx.redis.redis_pool import get_redis_client
 from onyx.redis.redis_pool import get_redis_replica_client
 from onyx.server.utils import make_short_id
+from onyx.utils.logger import format_error_for_logging
 from onyx.utils.logger import setup_logger
 
 logger = setup_logger()
@@ -119,11 +123,11 @@ def _is_external_group_sync_due(cc_pair: ConnectorCredentialPair) -> bool:
     soft_time_limit=JOB_TIMEOUT,
     bind=True,
 )
-def check_for_external_group_sync(self: Task, *, tenant_id: str | None) -> bool | None:
+def check_for_external_group_sync(self: Task, *, tenant_id: str) -> bool | None:
     # we need to use celery's redis client to access its redis data
     # (which lives on a different db number)
-    r = get_redis_client(tenant_id=tenant_id)
-    r_replica = get_redis_replica_client(tenant_id=tenant_id)
+    r = get_redis_client()
+    r_replica = get_redis_replica_client()
     r_celery: Redis = self.app.broker_connection().channel().client  # type: ignore
 
     lock_beat: RedisLock = r.lock(
@@ -140,7 +144,7 @@ def check_for_external_group_sync(self: Task, *, tenant_id: str | None) -> bool
 
     try:
         cc_pair_ids_to_sync: list[int] = []
-        with get_session_with_tenant(tenant_id) as db_session:
+        with get_session_with_current_tenant() as db_session:
             cc_pairs = get_all_auto_sync_cc_pairs(db_session)
 
             # We only want to sync one cc_pair per source type in
@@ -148,7 +152,10 @@ def check_for_external_group_sync(self: Task, *, tenant_id: str | None) -> bool
             for source in GROUP_PERMISSIONS_IS_CC_PAIR_AGNOSTIC:
                 # These are ordered by cc_pair id so the first one is the one we want
                 cc_pairs_to_dedupe = get_cc_pairs_by_source(
-                    db_session, source, only_sync=True
+                    db_session,
+                    source,
+                    access_type=AccessType.SYNC,
+                    status=ConnectorCredentialPairStatus.ACTIVE,
                 )
                 # We only want to sync one cc_pair per source type
                 # in GROUP_PERMISSIONS_IS_CC_PAIR_AGNOSTIC so we dedupe here
@@ -195,12 +202,17 @@ def check_for_external_group_sync(self: Task, *, tenant_id: str | None) -> bool
         task_logger.info(
             "Soft time limit exceeded, task is being terminated gracefully."
         )
-    except Exception:
+    except Exception as e:
+        error_msg = format_error_for_logging(e)
+        task_logger.warning(
+            f"Unexpected check_for_external_group_sync exception: tenant={tenant_id} {error_msg}"
+        )
         task_logger.exception(f"Unexpected exception: tenant={tenant_id}")
     finally:
         if lock_beat.owned():
             lock_beat.release()
 
+    task_logger.info(f"check_for_external_group_sync finished: tenant={tenant_id}")
     return True
 
 
@@ -208,7 +220,7 @@ def try_creating_external_group_sync_task(
     app: Celery,
     cc_pair_id: int,
     r: Redis,
-    tenant_id: str | None,
+    tenant_id: str,
 ) -> str | None:
     """Returns an int if syncing is needed. The int represents the number of sync tasks generated.
     Returns None if no syncing is required."""
@@ -230,7 +242,7 @@ def try_creating_external_group_sync_task(
         # create before setting fence to avoid race condition where the monitoring
         # task updates the sync record before it is created
         try:
-            with get_session_with_tenant(tenant_id) as db_session:
+            with get_session_with_current_tenant() as db_session:
                 insert_sync_record(
                     db_session=db_session,
                     entity_id=cc_pair_id,
@@ -267,12 +279,19 @@ def try_creating_external_group_sync_task(
         redis_connector.external_group_sync.set_fence(payload)
 
         payload_id = payload.id
-    except Exception:
+    except Exception as e:
+        error_msg = format_error_for_logging(e)
+        task_logger.warning(
+            f"Unexpected try_creating_external_group_sync_task exception: cc_pair={cc_pair_id} {error_msg}"
+        )
         task_logger.exception(
             f"Unexpected exception while trying to create external group sync task: cc_pair={cc_pair_id}"
         )
         return None
 
+    task_logger.info(
+        f"try_creating_external_group_sync_task finished: cc_pair={cc_pair_id} payload_id={payload_id}"
+    )
     return payload_id
 
 
@@ -287,7 +306,7 @@ def try_creating_external_group_sync_task(
 def connector_external_group_sync_generator_task(
     self: Task,
     cc_pair_id: int,
-    tenant_id: str | None,
+    tenant_id: str,
 ) -> None:
     """
     External group sync task for a given connector credential pair
@@ -296,7 +315,7 @@ def connector_external_group_sync_generator_task(
 
     redis_connector = RedisConnector(tenant_id, cc_pair_id)
 
-    r = get_redis_client(tenant_id=tenant_id)
+    r = get_redis_client()
 
     # this wait is needed to avoid a race condition where
     # the primary worker sends the task and it is immediately executed
@@ -357,16 +376,40 @@ def connector_external_group_sync_generator_task(
         payload.started = datetime.now(timezone.utc)
         redis_connector.external_group_sync.set_fence(payload)
 
-        with get_session_with_tenant(tenant_id) as db_session:
+        with get_session_with_current_tenant() as db_session:
             cc_pair = get_connector_credential_pair_from_id(
                 db_session=db_session,
                 cc_pair_id=cc_pair_id,
+                eager_load_credential=True,
             )
             if cc_pair is None:
                 raise ValueError(
                     f"No connector credential pair found for id: {cc_pair_id}"
                 )
 
+            try:
+                created = validate_ccpair_for_user(
+                    cc_pair.connector.id,
+                    cc_pair.credential.id,
+                    db_session,
+                    enforce_creation=False,
+                )
+                if not created:
+                    task_logger.warning(
+                        f"Unable to create connector credential pair for id: {cc_pair_id}"
+                    )
+            except Exception:
+                task_logger.exception(
+                    f"validate_ccpair_permissions_sync exceptioned: cc_pair={cc_pair_id}"
+                )
+                update_connector_credential_pair(
+                    db_session=db_session,
+                    connector_id=cc_pair.connector.id,
+                    credential_id=cc_pair.credential.id,
+                    status=ConnectorCredentialPairStatus.INVALID,
+                )
+                raise
+
             source_type = cc_pair.connector.source
 
             ext_group_sync_func = GROUP_PERMISSIONS_FUNC_MAP.get(source_type)
@@ -378,12 +421,23 @@ def connector_external_group_sync_generator_task(
             logger.info(
                 f"Syncing external groups for {source_type} for cc_pair: {cc_pair_id}"
             )
-
-            external_user_groups: list[ExternalUserGroup] = ext_group_sync_func(cc_pair)
+            external_user_groups: list[ExternalUserGroup] = []
+            try:
+                external_user_groups = ext_group_sync_func(tenant_id, cc_pair)
+            except ConnectorValidationError as e:
+                msg = f"Error syncing external groups for {source_type} for cc_pair: {cc_pair_id} {e}"
+                update_connector_credential_pair(
+                    db_session=db_session,
+                    connector_id=cc_pair.connector.id,
+                    credential_id=cc_pair.credential.id,
+                    status=ConnectorCredentialPairStatus.INVALID,
+                )
+                raise e
 
             logger.info(
                 f"Syncing {len(external_user_groups)} external user groups for {source_type}"
             )
+            logger.debug(f"New external user groups: {external_user_groups}")
 
             replace_user__ext_group_for_cc_pair(
                 db_session=db_session,
@@ -404,11 +458,19 @@ def connector_external_group_sync_generator_task(
                 sync_status=SyncStatus.SUCCESS,
             )
     except Exception as e:
+        error_msg = format_error_for_logging(e)
+        task_logger.warning(
+            f"External group sync exceptioned: cc_pair={cc_pair_id} payload_id={payload.id} {error_msg}"
+        )
+        task_logger.exception(
+            f"External group sync exceptioned: cc_pair={cc_pair_id} payload_id={payload.id}"
+        )
+
         msg = f"External group sync exceptioned: cc_pair={cc_pair_id} payload_id={payload.id}"
         task_logger.exception(msg)
         emit_background_error(msg + f"\n\n{e}", cc_pair_id=cc_pair_id)
 
-        with get_session_with_tenant(tenant_id) as db_session:
+        with get_session_with_current_tenant() as db_session:
             update_sync_record_status(
                 db_session=db_session,
                 entity_id=cc_pair_id,
@@ -431,7 +493,7 @@ def connector_external_group_sync_generator_task(
 
 
 def validate_external_group_sync_fences(
-    tenant_id: str | None,
+    tenant_id: str,
     celery_app: Celery,
     r: Redis,
     r_replica: Redis,
@@ -459,12 +521,11 @@ def validate_external_group_sync_fences(
         )
 
         lock_beat.reacquire()
-
     return
 
 
 def validate_external_group_sync_fence(
-    tenant_id: str | None,
+    tenant_id: str,
     key_bytes: bytes,
     reserved_tasks: set[str],
     r_celery: Redis,

backend/onyx/background/celery/tasks/indexing/utils.py

@@ -23,7 +23,7 @@ from onyx.configs.constants import OnyxCeleryQueues
 from onyx.configs.constants import OnyxCeleryTask
 from onyx.configs.constants import OnyxRedisConstants
 from onyx.db.engine import get_db_current_time
-from onyx.db.engine import get_session_with_tenant
+from onyx.db.engine import get_session_with_current_tenant
 from onyx.db.enums import ConnectorCredentialPairStatus
 from onyx.db.enums import IndexingStatus
 from onyx.db.enums import IndexModelStatus
@@ -93,27 +93,25 @@ def get_unfenced_index_attempt_ids(db_session: Session, r: redis.Redis) -> list[
     return unfenced_attempts
 
 
-class IndexingCallback(IndexingHeartbeatInterface):
+class IndexingCallbackBase(IndexingHeartbeatInterface):
     PARENT_CHECK_INTERVAL = 60
 
     def __init__(
         self,
         parent_pid: int,
         redis_connector: RedisConnector,
-        redis_connector_index: RedisConnectorIndex,
         redis_lock: RedisLock,
         redis_client: Redis,
     ):
         super().__init__()
         self.parent_pid = parent_pid
         self.redis_connector: RedisConnector = redis_connector
-        self.redis_connector_index: RedisConnectorIndex = redis_connector_index
         self.redis_lock: RedisLock = redis_lock
         self.redis_client = redis_client
         self.started: datetime = datetime.now(timezone.utc)
         self.redis_lock.reacquire()
 
-        self.last_tag: str = "IndexingCallback.__init__"
+        self.last_tag: str = f"{self.__class__.__name__}.__init__"
         self.last_lock_reacquire: datetime = datetime.now(timezone.utc)
         self.last_lock_monotonic = time.monotonic()
 
@@ -127,8 +125,8 @@ class IndexingCallback(IndexingHeartbeatInterface):
 
     def progress(self, tag: str, amount: int) -> None:
         # rkuo: this shouldn't be necessary yet because we spawn the process this runs inside
-        # with daemon = True. It seems likely some indexing tasks will need to spawn other processes eventually
-        # so leave this code in until we're ready to test it.
+        # with daemon=True. It seems likely some indexing tasks will need to spawn other processes
+        # eventually, which daemon=True prevents, so leave this code in until we're ready to test it.
 
         # if self.parent_pid:
         #     # check if the parent pid is alive so we aren't running as a zombie
@@ -143,8 +141,6 @@ class IndexingCallback(IndexingHeartbeatInterface):
         #         self.last_parent_check = now
 
         try:
-            self.redis_connector.prune.set_active()
-
             current_time = time.monotonic()
             if current_time - self.last_lock_monotonic >= (
                 CELERY_GENERIC_BEAT_LOCK_TIMEOUT / 4
@@ -156,7 +152,7 @@ class IndexingCallback(IndexingHeartbeatInterface):
             self.last_tag = tag
         except LockError:
             logger.exception(
-                f"IndexingCallback - lock.reacquire exceptioned: "
+                f"{self.__class__.__name__} - lock.reacquire exceptioned: "
                 f"lock_timeout={self.redis_lock.timeout} "
                 f"start={self.started} "
                 f"last_tag={self.last_tag} "
@@ -167,13 +163,31 @@ class IndexingCallback(IndexingHeartbeatInterface):
             redis_lock_dump(self.redis_lock, self.redis_client)
             raise
 
+
+class IndexingCallback(IndexingCallbackBase):
+    def __init__(
+        self,
+        parent_pid: int,
+        redis_connector: RedisConnector,
+        redis_lock: RedisLock,
+        redis_client: Redis,
+        redis_connector_index: RedisConnectorIndex,
+    ):
+        super().__init__(parent_pid, redis_connector, redis_lock, redis_client)
+
+        self.redis_connector_index: RedisConnectorIndex = redis_connector_index
+
+    def progress(self, tag: str, amount: int) -> None:
+        self.redis_connector_index.set_active()
+        self.redis_connector_index.set_connector_active()
+        super().progress(tag, amount)
         self.redis_client.incrby(
             self.redis_connector_index.generator_progress_key, amount
         )
 
 
 def validate_indexing_fence(
-    tenant_id: str | None,
+    tenant_id: str,
     key_bytes: bytes,
     reserved_tasks: set[str],
     r_celery: Redis,
@@ -240,7 +254,8 @@ def validate_indexing_fence(
         # it would be odd to get here as there isn't that much that can go wrong during
         # initial fence setup, but it's still worth making sure we can recover
         logger.info(
-            f"validate_indexing_fence - Resetting fence in basic state without any activity: fence={fence_key}"
+            f"validate_indexing_fence - "
+            f"Resetting fence in basic state without any activity: fence={fence_key}"
         )
         redis_connector_index.reset()
         return
@@ -296,7 +311,7 @@ def validate_indexing_fence(
 
 
 def validate_indexing_fences(
-    tenant_id: str | None,
+    tenant_id: str,
     r_replica: Redis,
     r_celery: Redis,
     lock_beat: RedisLock,
@@ -317,7 +332,7 @@ def validate_indexing_fences(
         if not key_str.startswith(RedisConnectorIndex.FENCE_PREFIX):
             continue
 
-        with get_session_with_tenant(tenant_id) as db_session:
+        with get_session_with_current_tenant() as db_session:
             validate_indexing_fence(
                 tenant_id,
                 key_bytes,
@@ -331,11 +346,10 @@ def validate_indexing_fences(
     return
 
 
-def _should_index(
+def should_index(
     cc_pair: ConnectorCredentialPair,
     last_index: IndexAttempt | None,
     search_settings_instance: SearchSettings,
-    search_settings_primary: bool,
     secondary_index_building: bool,
     db_session: Session,
 ) -> bool:
@@ -400,9 +414,9 @@ def _should_index(
     ):
         return False
 
-    if search_settings_primary:
+    if search_settings_instance.status.is_current():
         if cc_pair.indexing_trigger is not None:
-            # if a manual indexing trigger is on the cc pair, honor it for primary search settings
+            # if a manual indexing trigger is on the cc pair, honor it for live search settings
             return True
 
     # if no attempt has ever occurred, we should index regardless of refresh_freq
@@ -427,7 +441,7 @@ def try_creating_indexing_task(
     reindex: bool,
     db_session: Session,
     r: Redis,
-    tenant_id: str | None,
+    tenant_id: str,
 ) -> int | None:
     """Checks for any conditions that should block the indexing task from being
     created, then creates the task.

backend/onyx/background/celery/tasks/llm_model_update/tasks.py

@@ -8,7 +8,7 @@ from onyx.background.celery.apps.app_base import task_logger
 from onyx.configs.app_configs import JOB_TIMEOUT
 from onyx.configs.app_configs import LLM_MODEL_UPDATE_API_URL
 from onyx.configs.constants import OnyxCeleryTask
-from onyx.db.engine import get_session_with_tenant
+from onyx.db.engine import get_session_with_current_tenant
 from onyx.db.models import LLMProvider
 
 
@@ -59,7 +59,7 @@ def _process_model_list_response(model_list_json: Any) -> list[str]:
     trail=False,
     bind=True,
 )
-def check_for_llm_model_update(self: Task, *, tenant_id: str | None) -> bool | None:
+def check_for_llm_model_update(self: Task, *, tenant_id: str) -> bool | None:
     if not LLM_MODEL_UPDATE_API_URL:
         raise ValueError("LLM model update API URL not configured")
 
@@ -75,7 +75,7 @@ def check_for_llm_model_update(self: Task, *, tenant_id: str | None) -> bool | N
         return None
 
     # Then update the database with the fetched models
-    with get_session_with_tenant(tenant_id) as db_session:
+    with get_session_with_current_tenant() as db_session:
         # Get the default LLM provider
         default_provider = (
             db_session.query(LLMProvider)

backend/onyx/background/celery/tasks/monitoring/tasks.py

@@ -26,7 +26,8 @@ from onyx.configs.constants import OnyxCeleryTask
 from onyx.configs.constants import OnyxRedisLocks
 from onyx.db.engine import get_all_tenant_ids
 from onyx.db.engine import get_db_current_time
-from onyx.db.engine import get_session_with_tenant
+from onyx.db.engine import get_session_with_current_tenant
+from onyx.db.engine import get_session_with_shared_schema
 from onyx.db.enums import IndexingStatus
 from onyx.db.enums import SyncStatus
 from onyx.db.enums import SyncType
@@ -42,7 +43,6 @@ from onyx.utils.telemetry import optional_telemetry
 from onyx.utils.telemetry import RecordType
 from shared_configs.contextvars import CURRENT_TENANT_ID_CONTEXTVAR
 
-
 _MONITORING_SOFT_TIME_LIMIT = 60 * 5  # 5 minutes
 _MONITORING_TIME_LIMIT = _MONITORING_SOFT_TIME_LIMIT + 60  # 6 minutes
 
@@ -91,7 +91,7 @@ class Metric(BaseModel):
         }
         task_logger.info(json.dumps(data))
 
-    def emit(self, tenant_id: str | None) -> None:
+    def emit(self, tenant_id: str) -> None:
         # Convert value to appropriate type based on the input value
         bool_value = None
         float_value = None
@@ -656,7 +656,7 @@ def build_job_id(
     queue=OnyxCeleryQueues.MONITORING,
     bind=True,
 )
-def monitor_background_processes(self: Task, *, tenant_id: str | None) -> None:
+def monitor_background_processes(self: Task, *, tenant_id: str) -> None:
     """Collect and emit metrics about background processes.
     This task runs periodically to gather metrics about:
     - Queue lengths for different Celery queues
@@ -668,7 +668,7 @@ def monitor_background_processes(self: Task, *, tenant_id: str | None) -> None:
         CURRENT_TENANT_ID_CONTEXTVAR.set(tenant_id)
 
     task_logger.info("Starting background monitoring")
-    r = get_redis_client(tenant_id=tenant_id)
+    r = get_redis_client()
 
     lock_monitoring: RedisLock = r.lock(
         OnyxRedisLocks.MONITOR_BACKGROUND_PROCESSES_LOCK,
@@ -683,7 +683,7 @@ def monitor_background_processes(self: Task, *, tenant_id: str | None) -> None:
     try:
         # Get Redis client for Celery broker
         redis_celery = self.app.broker_connection().channel().client  # type: ignore
-        redis_std = get_redis_client(tenant_id=tenant_id)
+        redis_std = get_redis_client()
 
         # Define metric collection functions and their dependencies
         metric_functions: list[Callable[[], list[Metric]]] = [
@@ -693,7 +693,7 @@ def monitor_background_processes(self: Task, *, tenant_id: str | None) -> None:
         ]
 
         # Collect and log each metric
-        with get_session_with_tenant(tenant_id) as db_session:
+        with get_session_with_current_tenant() as db_session:
             for metric_fn in metric_functions:
                 metrics = metric_fn()
                 for metric in metrics:
@@ -771,12 +771,11 @@ def cloud_check_alembic() -> bool | None:
             if tenant_id is None:
                 continue
 
-            with get_session_with_tenant(tenant_id=None) as session:
+            with get_session_with_shared_schema() as session:
                 try:
                     result = session.execute(
                         text(f'SELECT * FROM "{tenant_id}".alembic_version LIMIT 1')
                     )
-
                     result_scalar: str | None = result.scalar_one_or_none()
                     if result_scalar is None:
                         raise ValueError("Alembic version should not be None.")
@@ -865,7 +864,7 @@ def cloud_monitor_celery_queues(
 
 
 @shared_task(name=OnyxCeleryTask.MONITOR_CELERY_QUEUES, ignore_result=True, bind=True)
-def monitor_celery_queues(self: Task, *, tenant_id: str | None) -> None:
+def monitor_celery_queues(self: Task, *, tenant_id: str) -> None:
     return monitor_celery_queues_helper(self)
 
 

backend/onyx/background/celery/tasks/periodic/tasks.py

@@ -15,7 +15,7 @@ from onyx.background.celery.apps.app_base import task_logger
 from onyx.configs.app_configs import JOB_TIMEOUT
 from onyx.configs.constants import OnyxCeleryTask
 from onyx.configs.constants import PostgresAdvisoryLocks
-from onyx.db.engine import get_session_with_tenant
+from onyx.db.engine import get_session_with_current_tenant
 
 
 @shared_task(
@@ -24,7 +24,7 @@ from onyx.db.engine import get_session_with_tenant
     bind=True,
     base=AbortableTask,
 )
-def kombu_message_cleanup_task(self: Any, tenant_id: str | None) -> int:
+def kombu_message_cleanup_task(self: Any, tenant_id: str) -> int:
     """Runs periodically to clean up the kombu_message table"""
 
     # we will select messages older than this amount to clean up
@@ -36,7 +36,7 @@ def kombu_message_cleanup_task(self: Any, tenant_id: str | None) -> int:
     ctx["deleted"] = 0
     ctx["cleanup_age"] = KOMBU_MESSAGE_CLEANUP_AGE
     ctx["page_limit"] = KOMBU_MESSAGE_CLEANUP_PAGE_LIMIT
-    with get_session_with_tenant(tenant_id) as db_session:
+    with get_session_with_current_tenant() as db_session:
         # Exit the task if we can't take the advisory lock
         result = db_session.execute(
             text("SELECT pg_try_advisory_lock(:id)"),

backend/onyx/background/celery/tasks/pruning/tasks.py

@@ -21,7 +21,7 @@ from onyx.background.celery.celery_redis import celery_get_queue_length
 from onyx.background.celery.celery_redis import celery_get_queued_task_ids
 from onyx.background.celery.celery_redis import celery_get_unacked_task_ids
 from onyx.background.celery.celery_utils import extract_ids_from_runnable_connector
-from onyx.background.celery.tasks.indexing.utils import IndexingCallback
+from onyx.background.celery.tasks.indexing.utils import IndexingCallbackBase
 from onyx.configs.app_configs import ALLOW_SIMULTANEOUS_PRUNING
 from onyx.configs.app_configs import JOB_TIMEOUT
 from onyx.configs.constants import CELERY_GENERIC_BEAT_LOCK_TIMEOUT
@@ -41,7 +41,7 @@ from onyx.db.connector_credential_pair import get_connector_credential_pair
 from onyx.db.connector_credential_pair import get_connector_credential_pair_from_id
 from onyx.db.connector_credential_pair import get_connector_credential_pairs
 from onyx.db.document import get_documents_for_connector_credential_pair
-from onyx.db.engine import get_session_with_tenant
+from onyx.db.engine import get_session_with_current_tenant
 from onyx.db.enums import ConnectorCredentialPairStatus
 from onyx.db.enums import SyncStatus
 from onyx.db.enums import SyncType
@@ -55,6 +55,7 @@ from onyx.redis.redis_connector_prune import RedisConnectorPrunePayload
 from onyx.redis.redis_pool import get_redis_client
 from onyx.redis.redis_pool import get_redis_replica_client
 from onyx.server.utils import make_short_id
+from onyx.utils.logger import format_error_for_logging
 from onyx.utils.logger import LoggerContextVars
 from onyx.utils.logger import pruning_ctx
 from onyx.utils.logger import setup_logger
@@ -62,6 +63,12 @@ from onyx.utils.logger import setup_logger
 logger = setup_logger()
 
 
+class PruneCallback(IndexingCallbackBase):
+    def progress(self, tag: str, amount: int) -> None:
+        self.redis_connector.prune.set_active()
+        super().progress(tag, amount)
+
+
 """Jobs / utils for kicking off pruning tasks."""
 
 
@@ -107,9 +114,9 @@ def _is_pruning_due(cc_pair: ConnectorCredentialPair) -> bool:
     soft_time_limit=JOB_TIMEOUT,
     bind=True,
 )
-def check_for_pruning(self: Task, *, tenant_id: str | None) -> bool | None:
-    r = get_redis_client(tenant_id=tenant_id)
-    r_replica = get_redis_replica_client(tenant_id=tenant_id)
+def check_for_pruning(self: Task, *, tenant_id: str) -> bool | None:
+    r = get_redis_client()
+    r_replica = get_redis_replica_client()
     r_celery: Redis = self.app.broker_connection().channel().client  # type: ignore
 
     lock_beat: RedisLock = r.lock(
@@ -127,14 +134,14 @@ def check_for_pruning(self: Task, *, tenant_id: str | None) -> bool | None:
         # but pruning only kicks off once per hour
         if not r.exists(OnyxRedisSignals.BLOCK_PRUNING):
             cc_pair_ids: list[int] = []
-            with get_session_with_tenant(tenant_id) as db_session:
+            with get_session_with_current_tenant() as db_session:
                 cc_pairs = get_connector_credential_pairs(db_session)
                 for cc_pair_entry in cc_pairs:
                     cc_pair_ids.append(cc_pair_entry.id)
 
             for cc_pair_id in cc_pair_ids:
                 lock_beat.reacquire()
-                with get_session_with_tenant(tenant_id) as db_session:
+                with get_session_with_current_tenant() as db_session:
                     cc_pair = get_connector_credential_pair_from_id(
                         db_session=db_session,
                         cc_pair_id=cc_pair_id,
@@ -182,18 +189,20 @@ def check_for_pruning(self: Task, *, tenant_id: str | None) -> bool | None:
 
             key_str = key_bytes.decode("utf-8")
             if key_str.startswith(RedisConnectorPrune.FENCE_PREFIX):
-                with get_session_with_tenant(tenant_id) as db_session:
+                with get_session_with_current_tenant() as db_session:
                     monitor_ccpair_pruning_taskset(tenant_id, key_bytes, r, db_session)
     except SoftTimeLimitExceeded:
         task_logger.info(
             "Soft time limit exceeded, task is being terminated gracefully."
         )
-    except Exception:
+    except Exception as e:
+        error_msg = format_error_for_logging(e)
+        task_logger.warning(f"Unexpected pruning check exception: {error_msg}")
         task_logger.exception("Unexpected exception during pruning check")
     finally:
         if lock_beat.owned():
             lock_beat.release()
-
+    task_logger.info(f"check_for_pruning finished: tenant={tenant_id}")
     return True
 
 
@@ -202,7 +211,7 @@ def try_creating_prune_generator_task(
     cc_pair: ConnectorCredentialPair,
     db_session: Session,
     r: Redis,
-    tenant_id: str | None,
+    tenant_id: str,
 ) -> str | None:
     """Checks for any conditions that should block the pruning generator task from being
     created, then creates the task.
@@ -295,13 +304,19 @@ def try_creating_prune_generator_task(
         redis_connector.prune.set_fence(payload)
 
         payload_id = payload.id
-    except Exception:
+    except Exception as e:
+        error_msg = format_error_for_logging(e)
+        task_logger.warning(
+            f"Unexpected try_creating_prune_generator_task exception: cc_pair={cc_pair.id} {error_msg}"
+        )
         task_logger.exception(f"Unexpected exception: cc_pair={cc_pair.id}")
         return None
     finally:
         if lock.owned():
             lock.release()
-
+    task_logger.info(
+        f"try_creating_prune_generator_task finished: cc_pair={cc_pair.id} payload_id={payload_id}"
+    )
     return payload_id
 
 
@@ -318,7 +333,7 @@ def connector_pruning_generator_task(
     cc_pair_id: int,
     connector_id: int,
     credential_id: int,
-    tenant_id: str | None,
+    tenant_id: str,
 ) -> None:
     """connector pruning task. For a cc pair, this task pulls all document IDs from the source
     and compares those IDs to locally stored documents and deletes all locally stored IDs missing
@@ -337,7 +352,7 @@ def connector_pruning_generator_task(
 
     redis_connector = RedisConnector(tenant_id, cc_pair_id)
 
-    r = get_redis_client(tenant_id=tenant_id)
+    r = get_redis_client()
 
     # this wait is needed to avoid a race condition where
     # the primary worker sends the task and it is immediately executed
@@ -395,7 +410,7 @@ def connector_pruning_generator_task(
         return None
 
     try:
-        with get_session_with_tenant(tenant_id) as db_session:
+        with get_session_with_current_tenant() as db_session:
             cc_pair = get_connector_credential_pair(
                 db_session=db_session,
                 connector_id=connector_id,
@@ -425,6 +440,7 @@ def connector_pruning_generator_task(
                 f"cc_pair={cc_pair_id} "
                 f"connector_source={cc_pair.connector.source}"
             )
+
             runnable_connector = instantiate_connector(
                 db_session,
                 cc_pair.connector.source,
@@ -434,12 +450,11 @@ def connector_pruning_generator_task(
             )
 
             search_settings = get_current_search_settings(db_session)
-            redis_connector_index = redis_connector.new_index(search_settings.id)
+            redis_connector.new_index(search_settings.id)
 
-            callback = IndexingCallback(
+            callback = PruneCallback(
                 0,
                 redis_connector,
-                redis_connector_index,
                 lock,
                 r,
             )
@@ -506,7 +521,7 @@ def connector_pruning_generator_task(
 
 
 def monitor_ccpair_pruning_taskset(
-    tenant_id: str | None, key_bytes: bytes, r: Redis, db_session: Session
+    tenant_id: str, key_bytes: bytes, r: Redis, db_session: Session
 ) -> None:
     fence_key = key_bytes.decode("utf-8")
     cc_pair_id_str = RedisConnector.get_id_from_fence_key(fence_key)
@@ -552,7 +567,7 @@ def monitor_ccpair_pruning_taskset(
 
 
 def validate_pruning_fences(
-    tenant_id: str | None,
+    tenant_id: str,
     r: Redis,
     r_replica: Redis,
     r_celery: Redis,
@@ -600,7 +615,7 @@ def validate_pruning_fences(
 
 
 def validate_pruning_fence(
-    tenant_id: str | None,
+    tenant_id: str,
     key_bytes: bytes,
     reserved_tasks: set[str],
     queued_tasks: set[str],

backend/onyx/background/celery/tasks/shared/RetryDocumentIndex.py

@@ -32,7 +32,7 @@ class RetryDocumentIndex:
         self,
         doc_id: str,
         *,
-        tenant_id: str | None,
+        tenant_id: str,
         chunk_count: int | None,
     ) -> int:
         return self.index.delete_single(
@@ -50,7 +50,7 @@ class RetryDocumentIndex:
         self,
         doc_id: str,
         *,
-        tenant_id: str | None,
+        tenant_id: str,
         chunk_count: int | None,
         fields: VespaDocumentFields,
     ) -> int:

backend/onyx/background/celery/tasks/shared/tasks.py

@@ -1,4 +1,5 @@
 import time
+from enum import Enum
 from http import HTTPStatus
 
 import httpx
@@ -27,7 +28,7 @@ from onyx.db.document import mark_document_as_modified
 from onyx.db.document import mark_document_as_synced
 from onyx.db.document_set import fetch_document_sets_for_document
 from onyx.db.engine import get_all_tenant_ids
-from onyx.db.engine import get_session_with_tenant
+from onyx.db.engine import get_session_with_current_tenant
 from onyx.db.search_settings import get_active_search_settings
 from onyx.document_index.factory import get_default_document_index
 from onyx.document_index.interfaces import VespaDocumentFields
@@ -45,6 +46,24 @@ LIGHT_SOFT_TIME_LIMIT = 105
 LIGHT_TIME_LIMIT = LIGHT_SOFT_TIME_LIMIT + 15
 
 
+class OnyxCeleryTaskCompletionStatus(str, Enum):
+    """The different statuses the watchdog can finish with.
+
+    TODO: create broader success/failure/abort categories
+    """
+
+    UNDEFINED = "undefined"
+
+    SUCCEEDED = "succeeded"
+
+    SKIPPED = "skipped"
+
+    SOFT_TIME_LIMIT = "soft_time_limit"
+
+    NON_RETRYABLE_EXCEPTION = "non_retryable_exception"
+    RETRYABLE_EXCEPTION = "retryable_exception"
+
+
 @shared_task(
     name=OnyxCeleryTask.DOCUMENT_BY_CC_PAIR_CLEANUP_TASK,
     soft_time_limit=LIGHT_SOFT_TIME_LIMIT,
@@ -57,7 +76,7 @@ def document_by_cc_pair_cleanup_task(
     document_id: str,
     connector_id: int,
     credential_id: int,
-    tenant_id: str | None,
+    tenant_id: str,
 ) -> bool:
     """A lightweight subtask used to clean up document to cc pair relationships.
     Created by connection deletion and connector pruning parent tasks."""
@@ -78,8 +97,10 @@ def document_by_cc_pair_cleanup_task(
 
     start = time.monotonic()
 
+    completion_status = OnyxCeleryTaskCompletionStatus.UNDEFINED
+
     try:
-        with get_session_with_tenant(tenant_id) as db_session:
+        with get_session_with_current_tenant() as db_session:
             action = "skip"
             chunks_affected = 0
 
@@ -105,10 +126,14 @@ def document_by_cc_pair_cleanup_task(
                     tenant_id=tenant_id,
                     chunk_count=chunk_count,
                 )
+
                 delete_documents_complete__no_commit(
                     db_session=db_session,
                     document_ids=[document_id],
                 )
+                db_session.commit()
+
+                completion_status = OnyxCeleryTaskCompletionStatus.SUCCEEDED
             elif count > 1:
                 action = "update"
 
@@ -152,10 +177,11 @@ def document_by_cc_pair_cleanup_task(
                 )
 
                 mark_document_as_synced(document_id, db_session)
-            else:
-                pass
+                db_session.commit()
 
-            db_session.commit()
+                completion_status = OnyxCeleryTaskCompletionStatus.SUCCEEDED
+            else:
+                completion_status = OnyxCeleryTaskCompletionStatus.SKIPPED
 
             elapsed = time.monotonic() - start
             task_logger.info(
@@ -167,57 +193,79 @@ def document_by_cc_pair_cleanup_task(
             )
     except SoftTimeLimitExceeded:
         task_logger.info(f"SoftTimeLimitExceeded exception. doc={document_id}")
-        return False
+        completion_status = OnyxCeleryTaskCompletionStatus.SOFT_TIME_LIMIT
     except Exception as ex:
         e: Exception | None = None
-        if isinstance(ex, RetryError):
-            task_logger.warning(
-                f"Tenacity retry failed: num_attempts={ex.last_attempt.attempt_number}"
+        while True:
+            if isinstance(ex, RetryError):
+                task_logger.warning(
+                    f"Tenacity retry failed: num_attempts={ex.last_attempt.attempt_number}"
+                )
+
+                # only set the inner exception if it is of type Exception
+                e_temp = ex.last_attempt.exception()
+                if isinstance(e_temp, Exception):
+                    e = e_temp
+            else:
+                e = ex
+
+            if isinstance(e, httpx.HTTPStatusError):
+                if e.response.status_code == HTTPStatus.BAD_REQUEST:
+                    task_logger.exception(
+                        f"Non-retryable HTTPStatusError: "
+                        f"doc={document_id} "
+                        f"status={e.response.status_code}"
+                    )
+                completion_status = (
+                    OnyxCeleryTaskCompletionStatus.NON_RETRYABLE_EXCEPTION
+                )
+                break
+
+            task_logger.exception(
+                f"document_by_cc_pair_cleanup_task exceptioned: doc={document_id}"
             )
 
-            # only set the inner exception if it is of type Exception
-            e_temp = ex.last_attempt.exception()
-            if isinstance(e_temp, Exception):
-                e = e_temp
-        else:
-            e = ex
-
-        if isinstance(e, httpx.HTTPStatusError):
-            if e.response.status_code == HTTPStatus.BAD_REQUEST:
-                task_logger.exception(
-                    f"Non-retryable HTTPStatusError: "
-                    f"doc={document_id} "
-                    f"status={e.response.status_code}"
+            completion_status = OnyxCeleryTaskCompletionStatus.RETRYABLE_EXCEPTION
+            if (
+                self.max_retries is not None
+                and self.request.retries >= self.max_retries
+            ):
+                # This is the last attempt! mark the document as dirty in the db so that it
+                # eventually gets fixed out of band via stale document reconciliation
+                task_logger.warning(
+                    f"Max celery task retries reached. Marking doc as dirty for reconciliation: "
+                    f"doc={document_id}"
                 )
-            return False
+                with get_session_with_current_tenant() as db_session:
+                    # delete the cc pair relationship now and let reconciliation clean it up
+                    # in vespa
+                    delete_document_by_connector_credential_pair__no_commit(
+                        db_session=db_session,
+                        document_id=document_id,
+                        connector_credential_pair_identifier=ConnectorCredentialPairIdentifier(
+                            connector_id=connector_id,
+                            credential_id=credential_id,
+                        ),
+                    )
+                    mark_document_as_modified(document_id, db_session)
+                completion_status = (
+                    OnyxCeleryTaskCompletionStatus.NON_RETRYABLE_EXCEPTION
+                )
+                break
 
-        task_logger.exception(f"Unexpected exception: doc={document_id}")
-
-        if self.request.retries < DOCUMENT_BY_CC_PAIR_CLEANUP_MAX_RETRIES:
-            # Still retrying. Exponential backoff from 2^4 to 2^6 ... i.e. 16, 32, 64
+            # Exponential backoff from 2^4 to 2^6 ... i.e. 16, 32, 64
             countdown = 2 ** (self.request.retries + 4)
-            self.retry(exc=e, countdown=countdown)
-        else:
-            # This is the last attempt! mark the document as dirty in the db so that it
-            # eventually gets fixed out of band via stale document reconciliation
-            task_logger.warning(
-                f"Max celery task retries reached. Marking doc as dirty for reconciliation: "
-                f"doc={document_id}"
-            )
-            with get_session_with_tenant(tenant_id) as db_session:
-                # delete the cc pair relationship now and let reconciliation clean it up
-                # in vespa
-                delete_document_by_connector_credential_pair__no_commit(
-                    db_session=db_session,
-                    document_id=document_id,
-                    connector_credential_pair_identifier=ConnectorCredentialPairIdentifier(
-                        connector_id=connector_id,
-                        credential_id=credential_id,
-                    ),
-                )
-                mark_document_as_modified(document_id, db_session)
+            self.retry(exc=e, countdown=countdown)  # this will raise a celery exception
+            break  # we won't hit this, but it looks weird not to have it
+    finally:
+        task_logger.info(
+            f"document_by_cc_pair_cleanup_task completed: status={completion_status.value} doc={document_id}"
+        )
+
+    if completion_status != OnyxCeleryTaskCompletionStatus.SUCCEEDED:
         return False
 
+    task_logger.info(f"document_by_cc_pair_cleanup_task finished: doc={document_id}")
     return True
 
 
@@ -249,7 +297,8 @@ def cloud_beat_task_generator(
         return None
 
     last_lock_time = time.monotonic()
-    tenant_ids: list[str] | list[None] = []
+    tenant_ids: list[str] = []
+    num_processed_tenants = 0
 
     try:
         tenant_ids = get_all_tenant_ids()
@@ -277,6 +326,8 @@ def cloud_beat_task_generator(
                 expires=expires,
                 ignore_result=True,
             )
+
+            num_processed_tenants += 1
     except SoftTimeLimitExceeded:
         task_logger.info(
             "Soft time limit exceeded, task is being terminated gracefully."
@@ -296,6 +347,7 @@ def cloud_beat_task_generator(
     task_logger.info(
         f"cloud_beat_task_generator finished: "
         f"task={task_name} "
+        f"num_processed_tenants={num_processed_tenants} "
         f"num_tenants={len(tenant_ids)} "
         f"elapsed={time_elapsed:.2f}"
     )

backend/onyx/background/celery/tasks/vespa/tasks.py

@@ -19,6 +19,7 @@ from onyx.background.celery.apps.app_base import task_logger
 from onyx.background.celery.tasks.shared.RetryDocumentIndex import RetryDocumentIndex
 from onyx.background.celery.tasks.shared.tasks import LIGHT_SOFT_TIME_LIMIT
 from onyx.background.celery.tasks.shared.tasks import LIGHT_TIME_LIMIT
+from onyx.background.celery.tasks.shared.tasks import OnyxCeleryTaskCompletionStatus
 from onyx.configs.app_configs import JOB_TIMEOUT
 from onyx.configs.app_configs import VESPA_SYNC_MAX_TASKS
 from onyx.configs.constants import CELERY_VESPA_SYNC_BEAT_LOCK_TIMEOUT
@@ -34,7 +35,7 @@ from onyx.db.document_set import fetch_document_sets
 from onyx.db.document_set import fetch_document_sets_for_document
 from onyx.db.document_set import get_document_set_by_id
 from onyx.db.document_set import mark_document_set_as_synced
-from onyx.db.engine import get_session_with_tenant
+from onyx.db.engine import get_session_with_current_tenant
 from onyx.db.enums import SyncStatus
 from onyx.db.enums import SyncType
 from onyx.db.models import DocumentSet
@@ -75,7 +76,7 @@ logger = setup_logger()
     trail=False,
     bind=True,
 )
-def check_for_vespa_sync_task(self: Task, *, tenant_id: str | None) -> bool | None:
+def check_for_vespa_sync_task(self: Task, *, tenant_id: str) -> bool | None:
     """Runs periodically to check if any document needs syncing.
     Generates sets of tasks for Celery if syncing is needed."""
 
@@ -84,8 +85,8 @@ def check_for_vespa_sync_task(self: Task, *, tenant_id: str | None) -> bool | No
 
     time_start = time.monotonic()
 
-    r = get_redis_client(tenant_id=tenant_id)
-    r_replica = get_redis_replica_client(tenant_id=tenant_id)
+    r = get_redis_client()
+    r_replica = get_redis_replica_client()
 
     lock_beat: RedisLock = r.lock(
         OnyxRedisLocks.CHECK_VESPA_SYNC_BEAT_LOCK,
@@ -98,7 +99,7 @@ def check_for_vespa_sync_task(self: Task, *, tenant_id: str | None) -> bool | No
 
     try:
         # 1/3: KICKOFF
-        with get_session_with_tenant(tenant_id) as db_session:
+        with get_session_with_current_tenant() as db_session:
             try_generate_stale_document_sync_tasks(
                 self.app, VESPA_SYNC_MAX_TASKS, db_session, r, lock_beat, tenant_id
             )
@@ -106,7 +107,7 @@ def check_for_vespa_sync_task(self: Task, *, tenant_id: str | None) -> bool | No
         # region document set scan
         lock_beat.reacquire()
         document_set_ids: list[int] = []
-        with get_session_with_tenant(tenant_id) as db_session:
+        with get_session_with_current_tenant() as db_session:
             # check if any document sets are not synced
             document_set_info = fetch_document_sets(
                 user_id=None, db_session=db_session, include_outdated=True
@@ -117,7 +118,7 @@ def check_for_vespa_sync_task(self: Task, *, tenant_id: str | None) -> bool | No
 
         for document_set_id in document_set_ids:
             lock_beat.reacquire()
-            with get_session_with_tenant(tenant_id) as db_session:
+            with get_session_with_current_tenant() as db_session:
                 try_generate_document_set_sync_tasks(
                     self.app, document_set_id, db_session, r, lock_beat, tenant_id
                 )
@@ -136,7 +137,7 @@ def check_for_vespa_sync_task(self: Task, *, tenant_id: str | None) -> bool | No
                 pass
             else:
                 usergroup_ids: list[int] = []
-                with get_session_with_tenant(tenant_id) as db_session:
+                with get_session_with_current_tenant() as db_session:
                     user_groups = fetch_user_groups(
                         db_session=db_session, only_up_to_date=False
                     )
@@ -146,7 +147,7 @@ def check_for_vespa_sync_task(self: Task, *, tenant_id: str | None) -> bool | No
 
                 for usergroup_id in usergroup_ids:
                     lock_beat.reacquire()
-                    with get_session_with_tenant(tenant_id) as db_session:
+                    with get_session_with_current_tenant() as db_session:
                         try_generate_user_group_sync_tasks(
                             self.app, usergroup_id, db_session, r, lock_beat, tenant_id
                         )
@@ -167,7 +168,7 @@ def check_for_vespa_sync_task(self: Task, *, tenant_id: str | None) -> bool | No
             if key_str == RedisGlobalConnectorCredentialPair.FENCE_KEY:
                 monitor_connector_taskset(r)
             elif key_str.startswith(RedisDocumentSet.FENCE_PREFIX):
-                with get_session_with_tenant(tenant_id) as db_session:
+                with get_session_with_current_tenant() as db_session:
                     monitor_document_set_taskset(tenant_id, key_bytes, r, db_session)
             elif key_str.startswith(RedisUserGroup.FENCE_PREFIX):
                 monitor_usergroup_taskset = (
@@ -177,7 +178,7 @@ def check_for_vespa_sync_task(self: Task, *, tenant_id: str | None) -> bool | No
                         noop_fallback,
                     )
                 )
-                with get_session_with_tenant(tenant_id) as db_session:
+                with get_session_with_current_tenant() as db_session:
                     monitor_usergroup_taskset(tenant_id, key_bytes, r, db_session)
 
     except SoftTimeLimitExceeded:
@@ -207,7 +208,7 @@ def try_generate_stale_document_sync_tasks(
     db_session: Session,
     r: Redis,
     lock_beat: RedisLock,
-    tenant_id: str | None,
+    tenant_id: str,
 ) -> int | None:
     # the fence is up, do nothing
 
@@ -283,7 +284,7 @@ def try_generate_document_set_sync_tasks(
     db_session: Session,
     r: Redis,
     lock_beat: RedisLock,
-    tenant_id: str | None,
+    tenant_id: str,
 ) -> int | None:
     lock_beat.reacquire()
 
@@ -360,7 +361,7 @@ def try_generate_user_group_sync_tasks(
     db_session: Session,
     r: Redis,
     lock_beat: RedisLock,
-    tenant_id: str | None,
+    tenant_id: str,
 ) -> int | None:
     lock_beat.reacquire()
 
@@ -447,7 +448,7 @@ def monitor_connector_taskset(r: Redis) -> None:
 
 
 def monitor_document_set_taskset(
-    tenant_id: str | None, key_bytes: bytes, r: Redis, db_session: Session
+    tenant_id: str, key_bytes: bytes, r: Redis, db_session: Session
 ) -> None:
     fence_key = key_bytes.decode("utf-8")
     document_set_id_str = RedisDocumentSet.get_id_from_fence_key(fence_key)
@@ -522,13 +523,13 @@ def monitor_document_set_taskset(
     time_limit=LIGHT_TIME_LIMIT,
     max_retries=3,
 )
-def vespa_metadata_sync_task(
-    self: Task, document_id: str, tenant_id: str | None
-) -> bool:
+def vespa_metadata_sync_task(self: Task, document_id: str, *, tenant_id: str) -> bool:
     start = time.monotonic()
 
+    completion_status = OnyxCeleryTaskCompletionStatus.UNDEFINED
+
     try:
-        with get_session_with_tenant(tenant_id) as db_session:
+        with get_session_with_current_tenant() as db_session:
             active_search_settings = get_active_search_settings(db_session)
             doc_index = get_default_document_index(
                 search_settings=active_search_settings.primary,
@@ -540,75 +541,103 @@ def vespa_metadata_sync_task(
 
             doc = get_document(document_id, db_session)
             if not doc:
-                return False
+                elapsed = time.monotonic() - start
+                task_logger.info(
+                    f"doc={document_id} "
+                    f"action=no_operation "
+                    f"elapsed={elapsed:.2f}"
+                )
+                completion_status = OnyxCeleryTaskCompletionStatus.SKIPPED
+            else:
+                # document set sync
+                doc_sets = fetch_document_sets_for_document(document_id, db_session)
+                update_doc_sets: set[str] = set(doc_sets)
 
-            # document set sync
-            doc_sets = fetch_document_sets_for_document(document_id, db_session)
-            update_doc_sets: set[str] = set(doc_sets)
+                # User group sync
+                doc_access = get_access_for_document(
+                    document_id=document_id, db_session=db_session
+                )
 
-            # User group sync
-            doc_access = get_access_for_document(
-                document_id=document_id, db_session=db_session
-            )
+                fields = VespaDocumentFields(
+                    document_sets=update_doc_sets,
+                    access=doc_access,
+                    boost=doc.boost,
+                    hidden=doc.hidden,
+                )
 
-            fields = VespaDocumentFields(
-                document_sets=update_doc_sets,
-                access=doc_access,
-                boost=doc.boost,
-                hidden=doc.hidden,
-            )
+                # update Vespa. OK if doc doesn't exist. Raises exception otherwise.
+                chunks_affected = retry_index.update_single(
+                    document_id,
+                    tenant_id=tenant_id,
+                    chunk_count=doc.chunk_count,
+                    fields=fields,
+                )
 
-            # update Vespa. OK if doc doesn't exist. Raises exception otherwise.
-            chunks_affected = retry_index.update_single(
-                document_id,
-                tenant_id=tenant_id,
-                chunk_count=doc.chunk_count,
-                fields=fields,
-            )
+                # update db last. Worst case = we crash right before this and
+                # the sync might repeat again later
+                mark_document_as_synced(document_id, db_session)
 
-            # update db last. Worst case = we crash right before this and
-            # the sync might repeat again later
-            mark_document_as_synced(document_id, db_session)
-
-            elapsed = time.monotonic() - start
-            task_logger.info(
-                f"doc={document_id} "
-                f"action=sync "
-                f"chunks={chunks_affected} "
-                f"elapsed={elapsed:.2f}"
-            )
+                elapsed = time.monotonic() - start
+                task_logger.info(
+                    f"doc={document_id} "
+                    f"action=sync "
+                    f"chunks={chunks_affected} "
+                    f"elapsed={elapsed:.2f}"
+                )
+                completion_status = OnyxCeleryTaskCompletionStatus.SUCCEEDED
     except SoftTimeLimitExceeded:
         task_logger.info(f"SoftTimeLimitExceeded exception. doc={document_id}")
-        return False
+        completion_status = OnyxCeleryTaskCompletionStatus.SOFT_TIME_LIMIT
     except Exception as ex:
         e: Exception | None = None
-        if isinstance(ex, RetryError):
-            task_logger.warning(
-                f"Tenacity retry failed: num_attempts={ex.last_attempt.attempt_number}"
+        while True:
+            if isinstance(ex, RetryError):
+                task_logger.warning(
+                    f"Tenacity retry failed: num_attempts={ex.last_attempt.attempt_number}"
+                )
+
+                # only set the inner exception if it is of type Exception
+                e_temp = ex.last_attempt.exception()
+                if isinstance(e_temp, Exception):
+                    e = e_temp
+            else:
+                e = ex
+
+            if isinstance(e, httpx.HTTPStatusError):
+                if e.response.status_code == HTTPStatus.BAD_REQUEST:
+                    task_logger.exception(
+                        f"Non-retryable HTTPStatusError: "
+                        f"doc={document_id} "
+                        f"status={e.response.status_code}"
+                    )
+                completion_status = (
+                    OnyxCeleryTaskCompletionStatus.NON_RETRYABLE_EXCEPTION
+                )
+                break
+
+            task_logger.exception(
+                f"vespa_metadata_sync_task exceptioned: doc={document_id}"
             )
 
-            # only set the inner exception if it is of type Exception
-            e_temp = ex.last_attempt.exception()
-            if isinstance(e_temp, Exception):
-                e = e_temp
-        else:
-            e = ex
-
-        if isinstance(e, httpx.HTTPStatusError):
-            if e.response.status_code == HTTPStatus.BAD_REQUEST:
-                task_logger.exception(
-                    f"Non-retryable HTTPStatusError: "
-                    f"doc={document_id} "
-                    f"status={e.response.status_code}"
+            completion_status = OnyxCeleryTaskCompletionStatus.RETRYABLE_EXCEPTION
+            if (
+                self.max_retries is not None
+                and self.request.retries >= self.max_retries
+            ):
+                completion_status = (
+                    OnyxCeleryTaskCompletionStatus.NON_RETRYABLE_EXCEPTION
                 )
-            return False
 
-        task_logger.exception(
-            f"Unexpected exception during vespa metadata sync: doc={document_id}"
+            # Exponential backoff from 2^4 to 2^6 ... i.e. 16, 32, 64
+            countdown = 2 ** (self.request.retries + 4)
+            self.retry(exc=e, countdown=countdown)  # this will raise a celery exception
+            break  # we won't hit this, but it looks weird not to have it
+    finally:
+        task_logger.info(
+            f"vespa_metadata_sync_task completed: status={completion_status.value} doc={document_id}"
         )
 
-        # Exponential backoff from 2^4 to 2^6 ... i.e. 16, 32, 64
-        countdown = 2 ** (self.request.retries + 4)
-        self.retry(exc=e, countdown=countdown)
+    if completion_status != OnyxCeleryTaskCompletionStatus.SUCCEEDED:
+        return False
 
     return True