style(chat): apply black formatting to process_message.py

Old code used queue.Queue() (unbounded, blocking put). New code introduced
queue.Queue(maxsize=100) + put(timeout=3.0) + silent drop on queue.Full —
a regression in all three callsites:

- Emitter.emit(): data packets silently dropped on queue full
- _run_model exception path: model errors silently lost
- _run_model finally (_MODEL_DONE): if dropped, drain loop hangs forever
  (models_remaining never reaches 0)

Fix: remove maxsize, remove all timeout= arguments, remove all
except queue.Full handlers. The drain_done early-return in emit() is the
correct disconnect mechanism; queue backpressure is not needed.

Also adds _completion_done: bool type annotation and fixes the queue drain
comment (no longer unblocking timed-out puts — just releasing memory).

.git-blame-ignore-revs

@@ -6,3 +6,4 @@
 
 3134e5f840c12c8f32613ce520101a047c89dcc2  # refactor(whitespace): rm temporary react fragments (#7161)
 ed3f72bc75f3e3a9ae9e4d8cd38278f9c97e78b4  # refactor(whitespace): rm react fragment #7190
+7b927e79c25f4ddfd18a067f489e122acd2c89de  # chore(format): format files where `ruff` and `black` agree (#9339)

.github/actions/slack-notify/action.yml

@@ -10,6 +10,9 @@ inputs:
   failed-jobs:
     description: "Deprecated alias for details"
     required: false
+  mention:
+    description: "GitHub username to resolve to a Slack @-mention. Replaces {mention} in details."
+    required: false
   title:
     description: "Title for the notification"
     required: false
@@ -26,6 +29,7 @@ runs:
         SLACK_WEBHOOK_URL: ${{ inputs.webhook-url }}
         DETAILS: ${{ inputs.details }}
         FAILED_JOBS: ${{ inputs.failed-jobs }}
+        MENTION_USER: ${{ inputs.mention }}
         TITLE: ${{ inputs.title }}
         REF_NAME: ${{ inputs.ref-name }}
         REPO: ${{ github.repository }}
@@ -52,6 +56,27 @@ runs:
           DETAILS="$FAILED_JOBS"
         fi
 
+        # Resolve {mention} placeholder if a GitHub username was provided.
+        # Looks up the username in user-mappings.json (co-located with this action)
+        # and replaces {mention} with <@SLACK_ID> for a Slack @-mention.
+        # Falls back to the plain GitHub username if not found in the mapping.
+        if [ -n "$MENTION_USER" ]; then
+          MAPPINGS_FILE="${GITHUB_ACTION_PATH}/user-mappings.json"
+          slack_id="$(jq -r --arg gh "$MENTION_USER" 'to_entries[] | select(.value | ascii_downcase == ($gh | ascii_downcase)) | .key' "$MAPPINGS_FILE" 2>/dev/null | head -1)"
+
+          if [ -n "$slack_id" ]; then
+            mention_text="<@${slack_id}>"
+          else
+            mention_text="${MENTION_USER}"
+          fi
+
+          DETAILS="${DETAILS//\{mention\}/$mention_text}"
+          TITLE="${TITLE//\{mention\}/}"
+        else
+          DETAILS="${DETAILS//\{mention\}/}"
+          TITLE="${TITLE//\{mention\}/}"
+        fi
+
         normalize_multiline() {
           printf '%s' "$1" | awk 'BEGIN { ORS=""; first=1 } { if (!first) printf "\\n"; printf "%s", $0; first=0 }'
         }

.github/actions/slack-notify/user-mappings.json

@@ -0,0 +1,18 @@
+{
+  "U05SAGZPEA1": "yuhongsun96",
+  "U05SAH6UGUD": "Weves",
+  "U07PWEQB7A5": "evan-onyx",
+  "U07V1SM68KF": "joachim-danswer",
+  "U08JZ9N3QNN": "raunakab",
+  "U08L24NCLJE": "Subash-Mohan",
+  "U090B9M07B2": "wenxi-onyx",
+  "U094RASDP0Q": "duo-onyx",
+  "U096L8ZQ85B": "justin-tahara",
+  "U09AHV8UBQX": "jessicasingh7",
+  "U09KAL5T3C2": "nmgarza5",
+  "U09KPGVQ70R": "acaprau",
+  "U09QR8KTSJH": "rohoswagger",
+  "U09RB4NTXA4": "jmelahman",
+  "U0A6K9VCY6A": "Danelegend",
+  "U0AGC4KH71A": "Bo-Onyx"
+}

.github/workflows/deployment.yml

@@ -44,7 +44,7 @@ jobs:
           fetch-tags: true
 
       - name: Setup uv
-        uses: astral-sh/setup-uv@5a095e7a2014a4212f075830d4f7277575a9d098 # ratchet:astral-sh/setup-uv@v7
+        uses: astral-sh/setup-uv@37802adc94f370d6bfd71619e3f0bf239e1f3b78 # ratchet:astral-sh/setup-uv@v7
         with:
           version: "0.9.9"
           enable-cache: false
@@ -165,7 +165,7 @@ jobs:
           fetch-depth: 0
 
       - name: Setup uv
-        uses: astral-sh/setup-uv@5a095e7a2014a4212f075830d4f7277575a9d098 # ratchet:astral-sh/setup-uv@v7
+        uses: astral-sh/setup-uv@37802adc94f370d6bfd71619e3f0bf239e1f3b78 # ratchet:astral-sh/setup-uv@v7
         with:
           version: "0.9.9"
           # NOTE: This isn't caching much and zizmor suggests this could be poisoned, so disable.
@@ -307,7 +307,7 @@ jobs:
             xdg-utils
 
       - name: setup node
-        uses: actions/setup-node@6044e13b5dc448c55e2357c09f80417699197238 # ratchet:actions/setup-node@v6.2.0
+        uses: actions/setup-node@53b83947a5a98c8d113130e565377fae1a50d02f # ratchet:actions/setup-node@v6.3.0
         with:
           node-version: 24
           package-manager-cache: false
@@ -455,7 +455,7 @@ jobs:
 
       - name: Docker meta
         id: meta
-        uses: docker/metadata-action@c299e40c65443455700f0fdfc63efafe5b349051 # ratchet:docker/metadata-action@v5
+        uses: docker/metadata-action@030e881283bb7a6894de51c315a6bfe6a94e05cf # ratchet:docker/metadata-action@v6.0.0
         with:
           images: ${{ needs.determine-builds.outputs.is-test-run == 'true' && env.RUNS_ON_ECR_CACHE || env.REGISTRY_IMAGE }}
           flavor: |
@@ -529,7 +529,7 @@ jobs:
 
       - name: Docker meta
         id: meta
-        uses: docker/metadata-action@c299e40c65443455700f0fdfc63efafe5b349051 # ratchet:docker/metadata-action@v5
+        uses: docker/metadata-action@030e881283bb7a6894de51c315a6bfe6a94e05cf # ratchet:docker/metadata-action@v6.0.0
         with:
           images: ${{ needs.determine-builds.outputs.is-test-run == 'true' && env.RUNS_ON_ECR_CACHE || env.REGISTRY_IMAGE }}
           flavor: |
@@ -607,7 +607,7 @@ jobs:
 
       - name: Docker meta
         id: meta
-        uses: docker/metadata-action@c299e40c65443455700f0fdfc63efafe5b349051 # ratchet:docker/metadata-action@v5
+        uses: docker/metadata-action@030e881283bb7a6894de51c315a6bfe6a94e05cf # ratchet:docker/metadata-action@v6.0.0
         with:
           images: ${{ needs.determine-builds.outputs.is-test-run == 'true' && env.RUNS_ON_ECR_CACHE || env.REGISTRY_IMAGE }}
           flavor: |
@@ -615,6 +615,7 @@ jobs:
           tags: |
             type=raw,value=${{ needs.determine-builds.outputs.is-test-run == 'true' && format('web-{0}', needs.determine-builds.outputs.sanitized-tag) || github.ref_name }}
             type=raw,value=${{ needs.determine-builds.outputs.is-test-run != 'true' && needs.determine-builds.outputs.is-latest == 'true' && 'latest' || '' }}
+            type=raw,value=${{ needs.determine-builds.outputs.is-test-run != 'true' && needs.determine-builds.outputs.is-latest == 'true' && 'craft-latest' || '' }}
             type=raw,value=${{ needs.determine-builds.outputs.is-test-run != 'true' && env.EDGE_TAG == 'true' && 'edge' || '' }}
             type=raw,value=${{ needs.determine-builds.outputs.is-test-run != 'true' && needs.determine-builds.outputs.is-beta == 'true' && 'beta' || '' }}
 
@@ -668,7 +669,7 @@ jobs:
 
       - name: Docker meta
         id: meta
-        uses: docker/metadata-action@c299e40c65443455700f0fdfc63efafe5b349051 # ratchet:docker/metadata-action@v5
+        uses: docker/metadata-action@030e881283bb7a6894de51c315a6bfe6a94e05cf # ratchet:docker/metadata-action@v6.0.0
         with:
           images: ${{ needs.determine-builds.outputs.is-test-run == 'true' && env.RUNS_ON_ECR_CACHE || env.REGISTRY_IMAGE }}
           flavor: |
@@ -703,6 +704,9 @@ jobs:
             NEXT_PUBLIC_FORGOT_PASSWORD_ENABLED=true
             NEXT_PUBLIC_INCLUDE_ERROR_POPUP_SUPPORT_LINK=true
             NODE_OPTIONS=--max-old-space-size=8192
+            SENTRY_RELEASE=${{ github.sha }}
+          secrets: |
+            sentry_auth_token=${{ secrets.SENTRY_AUTH_TOKEN }}
           cache-from: |
             type=registry,ref=${{ env.RUNS_ON_ECR_CACHE }}:cloudweb-cache-amd64
             type=registry,ref=${{ env.REGISTRY_IMAGE }}:latest
@@ -750,7 +754,7 @@ jobs:
 
       - name: Docker meta
         id: meta
-        uses: docker/metadata-action@c299e40c65443455700f0fdfc63efafe5b349051 # ratchet:docker/metadata-action@v5
+        uses: docker/metadata-action@030e881283bb7a6894de51c315a6bfe6a94e05cf # ratchet:docker/metadata-action@v6.0.0
         with:
           images: ${{ needs.determine-builds.outputs.is-test-run == 'true' && env.RUNS_ON_ECR_CACHE || env.REGISTRY_IMAGE }}
           flavor: |
@@ -785,6 +789,9 @@ jobs:
             NEXT_PUBLIC_FORGOT_PASSWORD_ENABLED=true
             NEXT_PUBLIC_INCLUDE_ERROR_POPUP_SUPPORT_LINK=true
             NODE_OPTIONS=--max-old-space-size=8192
+            SENTRY_RELEASE=${{ github.sha }}
+          secrets: |
+            sentry_auth_token=${{ secrets.SENTRY_AUTH_TOKEN }}
           cache-from: |
             type=registry,ref=${{ env.RUNS_ON_ECR_CACHE }}:cloudweb-cache-arm64
             type=registry,ref=${{ env.REGISTRY_IMAGE }}:latest
@@ -836,7 +843,7 @@ jobs:
 
       - name: Docker meta
         id: meta
-        uses: docker/metadata-action@c299e40c65443455700f0fdfc63efafe5b349051 # ratchet:docker/metadata-action@v5
+        uses: docker/metadata-action@030e881283bb7a6894de51c315a6bfe6a94e05cf # ratchet:docker/metadata-action@v6.0.0
         with:
           images: ${{ needs.determine-builds.outputs.is-test-run == 'true' && env.RUNS_ON_ECR_CACHE || env.REGISTRY_IMAGE }}
           flavor: |
@@ -894,7 +901,7 @@ jobs:
 
       - name: Docker meta
         id: meta
-        uses: docker/metadata-action@c299e40c65443455700f0fdfc63efafe5b349051 # ratchet:docker/metadata-action@v5
+        uses: docker/metadata-action@030e881283bb7a6894de51c315a6bfe6a94e05cf # ratchet:docker/metadata-action@v6.0.0
         with:
           images: ${{ needs.determine-builds.outputs.is-test-run == 'true' && env.RUNS_ON_ECR_CACHE || env.REGISTRY_IMAGE }}
           flavor: |
@@ -967,7 +974,7 @@ jobs:
 
       - name: Docker meta
         id: meta
-        uses: docker/metadata-action@c299e40c65443455700f0fdfc63efafe5b349051 # ratchet:docker/metadata-action@v5
+        uses: docker/metadata-action@030e881283bb7a6894de51c315a6bfe6a94e05cf # ratchet:docker/metadata-action@v6.0.0
         with:
           images: ${{ needs.determine-builds.outputs.is-test-run == 'true' && env.RUNS_ON_ECR_CACHE || env.REGISTRY_IMAGE }}
           flavor: |
@@ -1044,7 +1051,7 @@ jobs:
 
       - name: Docker meta
         id: meta
-        uses: docker/metadata-action@c299e40c65443455700f0fdfc63efafe5b349051 # ratchet:docker/metadata-action@v5
+        uses: docker/metadata-action@030e881283bb7a6894de51c315a6bfe6a94e05cf # ratchet:docker/metadata-action@v6.0.0
         with:
           images: ${{ needs.determine-builds.outputs.is-test-run == 'true' && env.RUNS_ON_ECR_CACHE || env.REGISTRY_IMAGE }}
           flavor: |
@@ -1105,7 +1112,7 @@ jobs:
 
       - name: Docker meta
         id: meta
-        uses: docker/metadata-action@c299e40c65443455700f0fdfc63efafe5b349051 # ratchet:docker/metadata-action@v5
+        uses: docker/metadata-action@030e881283bb7a6894de51c315a6bfe6a94e05cf # ratchet:docker/metadata-action@v6.0.0
         with:
           images: ${{ env.REGISTRY_IMAGE }}
           flavor: |
@@ -1178,7 +1185,7 @@ jobs:
 
       - name: Docker meta
         id: meta
-        uses: docker/metadata-action@c299e40c65443455700f0fdfc63efafe5b349051 # ratchet:docker/metadata-action@v5
+        uses: docker/metadata-action@030e881283bb7a6894de51c315a6bfe6a94e05cf # ratchet:docker/metadata-action@v6.0.0
         with:
           images: ${{ env.REGISTRY_IMAGE }}
           flavor: |
@@ -1256,15 +1263,13 @@ jobs:
 
       - name: Docker meta
         id: meta
-        uses: docker/metadata-action@c299e40c65443455700f0fdfc63efafe5b349051 # ratchet:docker/metadata-action@v5
+        uses: docker/metadata-action@030e881283bb7a6894de51c315a6bfe6a94e05cf # ratchet:docker/metadata-action@v6.0.0
         with:
           images: ${{ env.REGISTRY_IMAGE }}
           flavor: |
             latest=false
           tags: |
             type=raw,value=craft-latest
-            # TODO: Consider aligning craft-latest tags with regular backend builds (e.g., latest, edge, beta)
-            # to keep tagging strategy consistent across all backend images
 
       - name: Create and push manifest
         env:
@@ -1317,7 +1322,7 @@ jobs:
 
       - name: Docker meta
         id: meta
-        uses: docker/metadata-action@c299e40c65443455700f0fdfc63efafe5b349051 # ratchet:docker/metadata-action@v5
+        uses: docker/metadata-action@030e881283bb7a6894de51c315a6bfe6a94e05cf # ratchet:docker/metadata-action@v6.0.0
         with:
           images: ${{ needs.determine-builds.outputs.is-test-run == 'true' && env.RUNS_ON_ECR_CACHE || env.REGISTRY_IMAGE }}
           flavor: |
@@ -1397,7 +1402,7 @@ jobs:
 
       - name: Docker meta
         id: meta
-        uses: docker/metadata-action@c299e40c65443455700f0fdfc63efafe5b349051 # ratchet:docker/metadata-action@v5
+        uses: docker/metadata-action@030e881283bb7a6894de51c315a6bfe6a94e05cf # ratchet:docker/metadata-action@v6.0.0
         with:
           images: ${{ needs.determine-builds.outputs.is-test-run == 'true' && env.RUNS_ON_ECR_CACHE || env.REGISTRY_IMAGE }}
           flavor: |
@@ -1480,7 +1485,7 @@ jobs:
 
       - name: Docker meta
         id: meta
-        uses: docker/metadata-action@c299e40c65443455700f0fdfc63efafe5b349051 # ratchet:docker/metadata-action@v5
+        uses: docker/metadata-action@030e881283bb7a6894de51c315a6bfe6a94e05cf # ratchet:docker/metadata-action@v6.0.0
         with:
           images: ${{ needs.determine-builds.outputs.is-test-run == 'true' && env.RUNS_ON_ECR_CACHE || env.REGISTRY_IMAGE }}
           flavor: |
@@ -1488,6 +1493,7 @@ jobs:
           tags: |
             type=raw,value=${{ needs.determine-builds.outputs.is-test-run == 'true' && format('model-server-{0}', needs.determine-builds.outputs.sanitized-tag) || github.ref_name }}
             type=raw,value=${{ needs.determine-builds.outputs.is-test-run != 'true' && needs.determine-builds.outputs.is-latest == 'true' && 'latest' || '' }}
+            type=raw,value=${{ needs.determine-builds.outputs.is-test-run != 'true' && needs.determine-builds.outputs.is-latest == 'true' && 'craft-latest' || '' }}
             type=raw,value=${{ needs.determine-builds.outputs.is-test-run != 'true' && env.EDGE_TAG == 'true' && 'edge' || '' }}
             type=raw,value=${{ needs.determine-builds.outputs.is-test-run != 'true' && needs.determine-builds.outputs.is-beta-standalone == 'true' && 'beta' || '' }}
 

.github/workflows/helm-chart-releases.yml

@@ -47,7 +47,8 @@ jobs:
           done
 
       - name: Publish Helm charts to gh-pages
-        uses: stefanprodan/helm-gh-pages@0ad2bb377311d61ac04ad9eb6f252fb68e207260 # ratchet:stefanprodan/helm-gh-pages@v1.7.0
+        # NOTE: HEAD of https://github.com/stefanprodan/helm-gh-pages/pull/43
+        uses: stefanprodan/helm-gh-pages@ad32ad3b8720abfeaac83532fd1e9bdfca5bbe27 # zizmor: ignore[impostor-commit]
         with:
           token: ${{ secrets.GITHUB_TOKEN }}
           charts_dir: deployment/helm/charts

.github/workflows/nightly-llm-provider-chat.yml

@@ -35,6 +35,7 @@ jobs:
     needs: [provider-chat-test]
     if: failure() && github.event_name == 'schedule'
     runs-on: ubuntu-slim
+    environment: ci-protected
     timeout-minutes: 5
     steps:
       - name: Checkout

.github/workflows/post-merge-beta-cherry-pick.yml

@@ -114,7 +114,7 @@ jobs:
           ref: main
 
       - name: Install the latest version of uv
-        uses: astral-sh/setup-uv@5a095e7a2014a4212f075830d4f7277575a9d098 # ratchet:astral-sh/setup-uv@v7
+        uses: astral-sh/setup-uv@37802adc94f370d6bfd71619e3f0bf239e1f3b78 # ratchet:astral-sh/setup-uv@v7
         with:
           enable-cache: false
           version: "0.9.9"
@@ -183,6 +183,7 @@ jobs:
       - cherry-pick-to-latest-release
     if: needs.resolve-cherry-pick-request.outputs.should_cherrypick == 'true' && needs.resolve-cherry-pick-request.result == 'success' && needs.cherry-pick-to-latest-release.result == 'success'
     runs-on: ubuntu-slim
+    environment: ci-protected
     timeout-minutes: 10
     steps:
       - name: Checkout
@@ -207,7 +208,7 @@ jobs:
           CHERRY_PICK_PR_URL: ${{ needs.cherry-pick-to-latest-release.outputs.cherry_pick_pr_url }}
         run: |
           source_pr_url="https://github.com/${GITHUB_REPOSITORY}/pull/${SOURCE_PR_NUMBER}"
-          details="*Cherry-pick PR opened successfully.*\\n• source PR: ${source_pr_url}"
+          details="*Cherry-pick PR opened successfully.*\\n• author: {mention}\\n• source PR: ${source_pr_url}"
           if [ -n "${CHERRY_PICK_PR_URL}" ]; then
             details="${details}\\n• cherry-pick PR: ${CHERRY_PICK_PR_URL}"
           fi
@@ -221,6 +222,7 @@ jobs:
         uses: ./.github/actions/slack-notify
         with:
           webhook-url: ${{ secrets.CHERRY_PICK_PRS_WEBHOOK }}
+          mention: ${{ needs.resolve-cherry-pick-request.outputs.merged_by }}
           details: ${{ steps.success-summary.outputs.details }}
           title: "✅ Automated Cherry-Pick PR Opened"
           ref-name: ${{ github.event.pull_request.base.ref }}
@@ -231,6 +233,7 @@ jobs:
       - cherry-pick-to-latest-release
     if: always() && needs.resolve-cherry-pick-request.outputs.should_cherrypick == 'true' && (needs.resolve-cherry-pick-request.result == 'failure' || needs.cherry-pick-to-latest-release.result == 'failure')
     runs-on: ubuntu-slim
+    environment: ci-protected
     timeout-minutes: 10
     steps:
       - name: Checkout
@@ -275,20 +278,21 @@ jobs:
           else
             failed_job_label="cherry-pick-to-latest-release"
           fi
-          failed_jobs="• ${failed_job_label}\\n• source PR: ${source_pr_url}\\n• reason: ${reason_text}"
+          details="• author: {mention}\\n• ${failed_job_label}\\n• source PR: ${source_pr_url}\\n• reason: ${reason_text}"
           if [ -n "${MERGE_COMMIT_SHA}" ]; then
-            failed_jobs="${failed_jobs}\\n• merge SHA: ${MERGE_COMMIT_SHA}"
+            details="${details}\\n• merge SHA: ${MERGE_COMMIT_SHA}"
           fi
           if [ -n "${details_excerpt}" ]; then
-            failed_jobs="${failed_jobs}\\n• excerpt: ${details_excerpt}"
+            details="${details}\\n• excerpt: ${details_excerpt}"
           fi
 
-          echo "jobs=${failed_jobs}" >> "$GITHUB_OUTPUT"
+          echo "details=${details}" >> "$GITHUB_OUTPUT"
 
       - name: Notify #cherry-pick-prs about cherry-pick failure
         uses: ./.github/actions/slack-notify
         with:
           webhook-url: ${{ secrets.CHERRY_PICK_PRS_WEBHOOK }}
-          details: ${{ steps.failure-summary.outputs.jobs }}
+          mention: ${{ needs.resolve-cherry-pick-request.outputs.merged_by }}
+          details: ${{ steps.failure-summary.outputs.details }}
           title: "🚨 Automated Cherry-Pick Failed"
           ref-name: ${{ github.event.pull_request.base.ref }}

.github/workflows/pr-desktop-build.yml

@@ -50,7 +50,7 @@ jobs:
           persist-credentials: false
 
       - name: Setup node
-        uses: actions/setup-node@6044e13b5dc448c55e2357c09f80417699197238
+        uses: actions/setup-node@53b83947a5a98c8d113130e565377fae1a50d02f
         with:
           node-version: 24
           cache: "npm" # zizmor: ignore[cache-poisoning]
@@ -63,7 +63,7 @@ jobs:
           targets: ${{ matrix.target }}
 
       - name: Cache Cargo registry and build
-        uses: actions/cache@cdf6c1fa76f9f475f3d7449005a359c84ca0f306 # zizmor: ignore[cache-poisoning]
+        uses: actions/cache@668228422ae6a00e4ad889ee87cd7109ec5666a7 # zizmor: ignore[cache-poisoning]
         with:
           path: |
             ~/.cargo/bin/
@@ -105,7 +105,7 @@ jobs:
 
       - name: Upload build artifacts
         if: always()
-        uses: actions/upload-artifact@b7c566a772e6b6bfb58ed0dc250532a479d7789f
+        uses: actions/upload-artifact@bbbca2ddaa5d8feaa63e36b76fdaad77386f024f
         with:
           name: desktop-build-${{ matrix.platform }}-${{ github.run_id }}
           path: |

.github/workflows/pr-external-dependency-unit-tests.yml

@@ -7,6 +7,15 @@ on:
   merge_group:
   pull_request:
     branches: [main]
+    paths:
+      - "backend/**"
+      - "pyproject.toml"
+      - "uv.lock"
+      - ".github/workflows/pr-external-dependency-unit-tests.yml"
+      - ".github/actions/setup-python-and-install-dependencies/**"
+      - ".github/actions/setup-playwright/**"
+      - "deployment/docker_compose/docker-compose.yml"
+      - "deployment/docker_compose/docker-compose.dev.yml"
   push:
     tags:
       - "v*.*.*"
@@ -174,7 +183,7 @@ jobs:
 
       - name: Upload Docker logs
         if: failure()
-        uses: actions/upload-artifact@b7c566a772e6b6bfb58ed0dc250532a479d7789f
+        uses: actions/upload-artifact@bbbca2ddaa5d8feaa63e36b76fdaad77386f024f
         with:
           name: docker-logs-${{ matrix.test-dir }}
           path: docker-logs/

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

@@ -25,7 +25,7 @@ jobs:
     outputs:
       modules: ${{ steps.set-modules.outputs.modules }}
     steps:
-      - uses: actions/checkout@8e8c483db84b4bee98b60c0593521ed34d9990e8
+      - uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd
         with:
           persist-credentials: false
       - id: set-modules
@@ -39,7 +39,7 @@ jobs:
       matrix:
         modules: ${{ fromJSON(needs.detect-modules.outputs.modules) }}
     steps:
-      - uses: actions/checkout@8e8c483db84b4bee98b60c0593521ed34d9990e8 # ratchet:actions/checkout@v6
+      - uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # ratchet:actions/checkout@v6
         with:
           persist-credentials: false
       - uses: actions/setup-go@4dc6199c7b1a012772edbd06daecab0f50c9053c # zizmor: ignore[cache-poisoning]

.github/workflows/pr-helm-chart-testing.yml

@@ -41,7 +41,7 @@ jobs:
           version: v3.19.0
 
       - name: Set up chart-testing
-        uses: helm/chart-testing-action@b5eebdd9998021f29756c53432f48dab66394810
+        uses: helm/chart-testing-action@2e2940618cb426dce2999631d543b53cdcfc8527
         with:
           uv_version: "0.9.9"
 

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

@@ -466,7 +466,7 @@ jobs:
 
       - name: Upload logs
         if: always()
-        uses: actions/upload-artifact@b7c566a772e6b6bfb58ed0dc250532a479d7789f
+        uses: actions/upload-artifact@bbbca2ddaa5d8feaa63e36b76fdaad77386f024f
         with:
           name: docker-all-logs-${{ matrix.edition }}-${{ matrix.test-dir.name }}
           path: ${{ github.workspace }}/docker-compose.log
@@ -587,7 +587,7 @@ jobs:
 
       - name: Upload logs (onyx-lite)
         if: always()
-        uses: actions/upload-artifact@b7c566a772e6b6bfb58ed0dc250532a479d7789f
+        uses: actions/upload-artifact@bbbca2ddaa5d8feaa63e36b76fdaad77386f024f
         with:
           name: docker-all-logs-onyx-lite
           path: ${{ github.workspace }}/docker-compose-onyx-lite.log
@@ -725,7 +725,7 @@ jobs:
 
       - name: Upload logs (multi-tenant)
         if: always()
-        uses: actions/upload-artifact@b7c566a772e6b6bfb58ed0dc250532a479d7789f
+        uses: actions/upload-artifact@bbbca2ddaa5d8feaa63e36b76fdaad77386f024f
         with:
           name: docker-all-logs-multitenant
           path: ${{ github.workspace }}/docker-compose-multitenant.log

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

@@ -28,7 +28,7 @@ jobs:
           persist-credentials: false
 
       - name: Setup node
-        uses: actions/setup-node@6044e13b5dc448c55e2357c09f80417699197238 # ratchet:actions/setup-node@v4
+        uses: actions/setup-node@53b83947a5a98c8d113130e565377fae1a50d02f # ratchet:actions/setup-node@v4
         with:
           node-version: 22
           cache: "npm" # zizmor: ignore[cache-poisoning] test-only workflow; no deploy artifacts
@@ -44,7 +44,7 @@ jobs:
 
       - name: Upload coverage reports
         if: always()
-        uses: actions/upload-artifact@b7c566a772e6b6bfb58ed0dc250532a479d7789f
+        uses: actions/upload-artifact@bbbca2ddaa5d8feaa63e36b76fdaad77386f024f
         with:
           name: jest-coverage-${{ github.run_id }}
           path: ./web/coverage

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

@@ -272,7 +272,7 @@ jobs:
 
       - name: Setup node
         # zizmor: ignore[cache-poisoning] ephemeral runners; no release artifacts
-        uses: actions/setup-node@6044e13b5dc448c55e2357c09f80417699197238 # ratchet:actions/setup-node@v4
+        uses: actions/setup-node@53b83947a5a98c8d113130e565377fae1a50d02f # ratchet:actions/setup-node@v4
         with:
           node-version: 22
           cache: "npm" # zizmor: ignore[cache-poisoning]
@@ -284,7 +284,7 @@ jobs:
 
       - name: Cache playwright cache
         # zizmor: ignore[cache-poisoning] ephemeral runners; no release artifacts
-        uses: runs-on/cache@50350ad4242587b6c8c2baa2e740b1bc11285ff4 # ratchet:runs-on/cache@v4
+        uses: runs-on/cache@a5f51d6f3fece787d03b7b4e981c82538a0654ed # ratchet:runs-on/cache@v4
         with:
           path: ~/.cache/ms-playwright
           key: ${{ runner.os }}-playwright-npm-${{ hashFiles('web/package-lock.json') }}
@@ -445,7 +445,7 @@ jobs:
         run: |
           npx playwright test --project ${PROJECT}
 
-      - uses: actions/upload-artifact@b7c566a772e6b6bfb58ed0dc250532a479d7789f
+      - uses: actions/upload-artifact@bbbca2ddaa5d8feaa63e36b76fdaad77386f024f
         if: always()
         with:
           # Includes test results and trace.zip files
@@ -454,7 +454,7 @@ jobs:
           retention-days: 30
 
       - name: Upload screenshots
-        uses: actions/upload-artifact@b7c566a772e6b6bfb58ed0dc250532a479d7789f
+        uses: actions/upload-artifact@bbbca2ddaa5d8feaa63e36b76fdaad77386f024f
         if: always()
         with:
           name: playwright-screenshots-${{ matrix.project }}-${{ github.run_id }}
@@ -471,7 +471,7 @@ jobs:
 
       - name: Install the latest version of uv
         if: always()
-        uses: astral-sh/setup-uv@5a095e7a2014a4212f075830d4f7277575a9d098 # ratchet:astral-sh/setup-uv@v7
+        uses: astral-sh/setup-uv@37802adc94f370d6bfd71619e3f0bf239e1f3b78 # ratchet:astral-sh/setup-uv@v7
         with:
           enable-cache: false
           version: "0.9.9"
@@ -534,7 +534,7 @@ jobs:
             "s3://${PLAYWRIGHT_S3_BUCKET}/reports/pr-${PR_NUMBER}/${RUN_ID}/${PROJECT}/"
 
       - name: Upload visual diff summary
-        uses: actions/upload-artifact@b7c566a772e6b6bfb58ed0dc250532a479d7789f
+        uses: actions/upload-artifact@bbbca2ddaa5d8feaa63e36b76fdaad77386f024f
         if: always()
         with:
           name: screenshot-diff-summary-${{ matrix.project }}
@@ -543,7 +543,7 @@ jobs:
           retention-days: 5
 
       - name: Upload visual diff report artifact
-        uses: actions/upload-artifact@b7c566a772e6b6bfb58ed0dc250532a479d7789f
+        uses: actions/upload-artifact@bbbca2ddaa5d8feaa63e36b76fdaad77386f024f
         if: always()
         with:
           name: screenshot-diff-report-${{ matrix.project }}-${{ github.run_id }}
@@ -590,7 +590,7 @@ jobs:
 
       - name: Upload logs
         if: success() || failure()
-        uses: actions/upload-artifact@b7c566a772e6b6bfb58ed0dc250532a479d7789f
+        uses: actions/upload-artifact@bbbca2ddaa5d8feaa63e36b76fdaad77386f024f
         with:
           name: docker-logs-${{ matrix.project }}-${{ github.run_id }}
           path: ${{ github.workspace }}/docker-compose.log
@@ -614,7 +614,7 @@ jobs:
 
       - name: Setup node
         # zizmor: ignore[cache-poisoning] ephemeral runners; no release artifacts
-        uses: actions/setup-node@6044e13b5dc448c55e2357c09f80417699197238 # ratchet:actions/setup-node@v4
+        uses: actions/setup-node@53b83947a5a98c8d113130e565377fae1a50d02f # ratchet:actions/setup-node@v4
         with:
           node-version: 22
           cache: "npm" # zizmor: ignore[cache-poisoning]
@@ -626,7 +626,7 @@ jobs:
 
       - name: Cache playwright cache
         # zizmor: ignore[cache-poisoning] ephemeral runners; no release artifacts
-        uses: runs-on/cache@50350ad4242587b6c8c2baa2e740b1bc11285ff4 # ratchet:runs-on/cache@v4
+        uses: runs-on/cache@a5f51d6f3fece787d03b7b4e981c82538a0654ed # ratchet:runs-on/cache@v4
         with:
           path: ~/.cache/ms-playwright
           key: ${{ runner.os }}-playwright-npm-${{ hashFiles('web/package-lock.json') }}
@@ -674,7 +674,7 @@ jobs:
         working-directory: ./web
         run: npx playwright test --project lite
 
-      - uses: actions/upload-artifact@b7c566a772e6b6bfb58ed0dc250532a479d7789f
+      - uses: actions/upload-artifact@bbbca2ddaa5d8feaa63e36b76fdaad77386f024f
         if: always()
         with:
           name: playwright-test-results-lite-${{ github.run_id }}
@@ -692,7 +692,7 @@ jobs:
 
       - name: Upload logs
         if: success() || failure()
-        uses: actions/upload-artifact@b7c566a772e6b6bfb58ed0dc250532a479d7789f
+        uses: actions/upload-artifact@bbbca2ddaa5d8feaa63e36b76fdaad77386f024f
         with:
           name: docker-logs-lite-${{ github.run_id }}
           path: ${{ github.workspace }}/docker-compose.log

.github/workflows/pr-python-checks.yml

@@ -56,7 +56,7 @@ jobs:
 
       - name: Cache mypy cache
         if: ${{ vars.DISABLE_MYPY_CACHE != 'true' }}
-        uses: runs-on/cache@50350ad4242587b6c8c2baa2e740b1bc11285ff4 # ratchet:runs-on/cache@v4
+        uses: runs-on/cache@a5f51d6f3fece787d03b7b4e981c82538a0654ed # ratchet:runs-on/cache@v4
         with:
           path: .mypy_cache
           key: mypy-${{ runner.os }}-${{ github.base_ref || github.event.merge_group.base_ref || 'main' }}-${{ hashFiles('**/*.py', '**/*.pyi', 'pyproject.toml') }}

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

@@ -7,6 +7,13 @@ on:
   merge_group:
   pull_request:
     branches: [main]
+    paths:
+      - "backend/**"
+      - "pyproject.toml"
+      - "uv.lock"
+      - ".github/workflows/pr-python-connector-tests.yml"
+      - ".github/actions/setup-python-and-install-dependencies/**"
+      - ".github/actions/setup-playwright/**"
   push:
     tags:
       - "v*.*.*"

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

@@ -31,6 +31,7 @@ jobs:
       - runner=4cpu-linux-arm64
       - "run-id=${{ github.run_id }}-model-check"
       - "extras=ecr-cache"
+    environment: ci-protected
     timeout-minutes: 45
 
     env:
@@ -73,7 +74,7 @@ jobs:
         uses: docker/setup-buildx-action@8d2750c68a42422c14e847fe6c8ac0403b4cbd6f
 
       - name: Build and load
-        uses: docker/bake-action@5be5f02ff8819ecd3092ea6b2e6261c31774f2b4 # ratchet:docker/bake-action@v6
+        uses: docker/bake-action@82490499d2e5613fcead7e128237ef0b0ea210f7 # ratchet:docker/bake-action@v7.0.0
         env:
           TAG: model-server-${{ github.run_id }}
         with:
@@ -122,7 +123,7 @@ jobs:
 
       - name: Upload logs
         if: always()
-        uses: actions/upload-artifact@b7c566a772e6b6bfb58ed0dc250532a479d7789f
+        uses: actions/upload-artifact@bbbca2ddaa5d8feaa63e36b76fdaad77386f024f
         with:
           name: docker-all-logs
           path: ${{ github.workspace }}/docker-compose.log

.github/workflows/pr-quality-checks.yml

@@ -30,7 +30,7 @@ jobs:
       - name: Setup Terraform
         uses: hashicorp/setup-terraform@5e8dbf3c6d9deaf4193ca7a8fb23f2ac83bb6c85 # ratchet:hashicorp/setup-terraform@v4.0.0
       - name: Setup node
-        uses: actions/setup-node@6044e13b5dc448c55e2357c09f80417699197238 # ratchet:actions/setup-node@v6
+        uses: actions/setup-node@53b83947a5a98c8d113130e565377fae1a50d02f # ratchet:actions/setup-node@v6
         with: # zizmor: ignore[cache-poisoning]
           node-version: 22
           cache: "npm"

.github/workflows/preview.yml

@@ -15,6 +15,7 @@ permissions:
 jobs:
   Deploy-Preview:
     runs-on: ubuntu-latest
+    environment: ci-protected
     timeout-minutes: 30
     steps:
       - uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd
@@ -22,7 +23,7 @@ jobs:
           persist-credentials: false
 
       - name: Setup node
-        uses: actions/setup-node@6044e13b5dc448c55e2357c09f80417699197238 # ratchet:actions/setup-node@v4
+        uses: actions/setup-node@53b83947a5a98c8d113130e565377fae1a50d02f # ratchet:actions/setup-node@v4
         with:
           node-version: 22
           cache: "npm"

.github/workflows/release-cli.yml

@@ -13,27 +13,20 @@ jobs:
     permissions:
       id-token: write
     timeout-minutes: 10
-    strategy:
-      matrix:
-        os-arch:
-          - { goos: "linux", goarch: "amd64" }
-          - { goos: "linux", goarch: "arm64" }
-          - { goos: "windows", goarch: "amd64" }
-          - { goos: "windows", goarch: "arm64" }
-          - { goos: "darwin", goarch: "amd64" }
-          - { goos: "darwin", goarch: "arm64" }
     steps:
       - uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # ratchet:actions/checkout@v6
         with:
           persist-credentials: false
-      - uses: astral-sh/setup-uv@5a095e7a2014a4212f075830d4f7277575a9d098 # ratchet:astral-sh/setup-uv@v7
+      - uses: astral-sh/setup-uv@37802adc94f370d6bfd71619e3f0bf239e1f3b78 # ratchet:astral-sh/setup-uv@v7
         with:
           enable-cache: false
           version: "0.9.9"
       - run: |
-          GOOS="${{ matrix.os-arch.goos }}" \
-          GOARCH="${{ matrix.os-arch.goarch }}" \
-          uv build --wheel
+          for goos in linux windows darwin; do
+            for goarch in amd64 arm64; do
+              GOOS="$goos" GOARCH="$goarch" uv build --wheel
+            done
+          done
         working-directory: cli
       - run: uv publish
         working-directory: cli

.github/workflows/release-devtools.yml

@@ -26,7 +26,7 @@ jobs:
       - uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # ratchet:actions/checkout@v6
         with:
           persist-credentials: false
-      - uses: astral-sh/setup-uv@5a095e7a2014a4212f075830d4f7277575a9d098 # ratchet:astral-sh/setup-uv@v7
+      - uses: astral-sh/setup-uv@37802adc94f370d6bfd71619e3f0bf239e1f3b78 # ratchet:astral-sh/setup-uv@v7
         with:
           enable-cache: false
           version: "0.9.9"

.github/workflows/reusable-nightly-llm-provider-chat.yml

@@ -319,7 +319,7 @@ jobs:
 
       - name: Upload logs
         if: always()
-        uses: actions/upload-artifact@b7c566a772e6b6bfb58ed0dc250532a479d7789f
+        uses: actions/upload-artifact@bbbca2ddaa5d8feaa63e36b76fdaad77386f024f
         with:
           name: docker-all-logs-nightly-${{ matrix.provider }}-llm-provider
           path: |

.github/workflows/sandbox-deployment.yml

@@ -125,7 +125,7 @@ jobs:
 
       - name: Docker meta
         id: meta
-        uses: docker/metadata-action@c299e40c65443455700f0fdfc63efafe5b349051 # ratchet:docker/metadata-action@v5
+        uses: docker/metadata-action@030e881283bb7a6894de51c315a6bfe6a94e05cf # ratchet:docker/metadata-action@v6.0.0
         with:
           images: ${{ env.REGISTRY_IMAGE }}
           flavor: |
@@ -195,7 +195,7 @@ jobs:
 
       - name: Docker meta
         id: meta
-        uses: docker/metadata-action@c299e40c65443455700f0fdfc63efafe5b349051 # ratchet:docker/metadata-action@v5
+        uses: docker/metadata-action@030e881283bb7a6894de51c315a6bfe6a94e05cf # ratchet:docker/metadata-action@v6.0.0
         with:
           images: ${{ env.REGISTRY_IMAGE }}
           flavor: |
@@ -268,7 +268,7 @@ jobs:
 
       - name: Docker meta
         id: meta
-        uses: docker/metadata-action@c299e40c65443455700f0fdfc63efafe5b349051 # ratchet:docker/metadata-action@v5
+        uses: docker/metadata-action@030e881283bb7a6894de51c315a6bfe6a94e05cf # ratchet:docker/metadata-action@v6.0.0
         with:
           images: ${{ env.REGISTRY_IMAGE }}
           flavor: |

.github/workflows/storybook-deploy.yml

@@ -25,6 +25,7 @@ permissions:
 jobs:
   Deploy-Storybook:
     runs-on: ubuntu-latest
+    environment: ci-protected
     timeout-minutes: 30
     steps:
       - uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # ratchet:actions/checkout@v4
@@ -32,7 +33,7 @@ jobs:
           persist-credentials: false
 
       - name: Setup node
-        uses: actions/setup-node@6044e13b5dc448c55e2357c09f80417699197238 # ratchet:actions/setup-node@v4
+        uses: actions/setup-node@53b83947a5a98c8d113130e565377fae1a50d02f # ratchet:actions/setup-node@v4
         with:
           node-version: 22
           cache: "npm"
@@ -54,6 +55,7 @@ jobs:
     needs: Deploy-Storybook
     if: always() && needs.Deploy-Storybook.result == 'failure'
     runs-on: ubuntu-latest
+    environment: ci-protected
     timeout-minutes: 10
     steps:
       - uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # ratchet:actions/checkout@v4

.github/workflows/sync_foss.yml

@@ -9,6 +9,7 @@ on:
 jobs:
   sync-foss:
     runs-on: ubuntu-latest
+    environment: ci-protected
     timeout-minutes: 45
     permissions:
       contents: read

.github/workflows/tag-nightly.yml

@@ -11,6 +11,7 @@ permissions:
 jobs:
   create-and-push-tag:
     runs-on: ubuntu-slim
+    environment: ci-protected
     timeout-minutes: 45
 
     steps:

.github/workflows/zizmor.yml

@@ -24,7 +24,7 @@ jobs:
           persist-credentials: false
 
       - name: Install the latest version of uv
-        uses: astral-sh/setup-uv@5a095e7a2014a4212f075830d4f7277575a9d098 # ratchet:astral-sh/setup-uv@v7
+        uses: astral-sh/setup-uv@37802adc94f370d6bfd71619e3f0bf239e1f3b78 # ratchet:astral-sh/setup-uv@v7
         with:
           enable-cache: false
           version: "0.9.9"

.greptile/config.json

@@ -0,0 +1,64 @@
+{
+    "labels": [],
+    "comment": "",
+    "fixWithAI": true,
+    "hideFooter": false,
+    "strictness": 3,
+    "statusCheck": true,
+    "commentTypes": [
+      "logic",
+      "syntax",
+      "style"
+    ],
+    "instructions": "",
+    "disabledLabels": [],
+    "excludeAuthors": [
+      "dependabot[bot]",
+      "renovate[bot]"
+    ],
+    "ignoreKeywords": "",
+    "ignorePatterns": "",
+    "includeAuthors": [],
+    "summarySection": {
+      "included": true,
+      "collapsible": false,
+      "defaultOpen": false
+    },
+    "excludeBranches": [],
+    "fileChangeLimit": 300,
+    "includeBranches": [],
+    "includeKeywords": "",
+    "triggerOnUpdates": true,
+    "updateExistingSummaryComment": true,
+    "updateSummaryOnly": false,
+    "issuesTableSection": {
+      "included": true,
+      "collapsible": false,
+      "defaultOpen": false
+    },
+    "statusCommentsEnabled": true,
+    "confidenceScoreSection": {
+      "included": true,
+      "collapsible": false
+    },
+    "sequenceDiagramSection": {
+      "included": true,
+      "collapsible": false,
+      "defaultOpen": false
+    },
+    "shouldUpdateDescription": false,
+    "rules": [
+      {
+        "scope": ["web/**"],
+        "rule": "In Onyx's Next.js app, the `app/ee/admin/` directory is a filesystem convention for Enterprise Edition route overrides — it does NOT add an `/ee/` prefix to the URL. Both `app/admin/groups/page.tsx` and `app/ee/admin/groups/page.tsx` serve the same URL `/admin/groups`. Hardcoded `/admin/...` paths in router.push() calls are correct and do NOT break EE deployments. Do not flag hardcoded admin paths as bugs."
+      },
+      {
+        "scope": ["web/**"],
+        "rule": "In Onyx, each API key creates a unique user row in the database with a unique `user_id` (UUID). There is a 1:1 mapping between API keys and their backing user records. Multiple API keys do NOT share the same `user_id`. Do not flag potential duplicate row IDs when using `user_id` from API key descriptors."
+      },
+      {
+        "scope": ["backend/**/*.py"],
+        "rule": "Never raise HTTPException directly in business code. Use `raise OnyxError(OnyxErrorCode.XXX, \"message\")` from `onyx.error_handling.exceptions`. A global FastAPI exception handler converts OnyxError into structured JSON responses with {\"error_code\": \"...\", \"detail\": \"...\"}. Error codes are defined in `onyx.error_handling.error_codes.OnyxErrorCode`. For upstream errors with dynamic HTTP status codes, use `status_code_override`: `raise OnyxError(OnyxErrorCode.BAD_GATEWAY, detail, status_code_override=upstream_status)`."
+      }
+    ]
+}

.greptile/files.json

@@ -0,0 +1,57 @@
+[
+  {
+    "scope": [],
+    "path": "contributing_guides/best_practices.md",
+    "description": "Best practices for contributing to the codebase"
+  },
+  {
+    "scope": ["web/**"],
+    "path": "web/AGENTS.md",
+    "description": "Frontend coding standards for the web directory"
+  },
+  {
+    "scope": ["web/**"],
+    "path": "web/tests/README.md",
+    "description": "Frontend testing guide and conventions"
+  },
+  {
+    "scope": ["web/**"],
+    "path": "web/CLAUDE.md",
+    "description": "Single source of truth for frontend coding standards"
+  },
+  {
+    "scope": ["web/**"],
+    "path": "web/lib/opal/README.md",
+    "description": "Opal component library usage guide"
+  },
+  {
+    "scope": ["backend/**"],
+    "path": "backend/tests/README.md",
+    "description": "Backend testing guide covering all 4 test types, fixtures, and conventions"
+  },
+  {
+    "scope": ["backend/onyx/connectors/**"],
+    "path": "backend/onyx/connectors/README.md",
+    "description": "Connector development guide covering design, interfaces, and required changes"
+  },
+  {
+    "scope": [],
+    "path": "CLAUDE.md",
+    "description": "Project instructions and coding standards"
+  },
+  {
+    "scope": [],
+    "path": "backend/alembic/README.md",
+    "description": "Migration guidance, including multi-tenant migration behavior"
+  },
+  {
+    "scope": [],
+    "path": "deployment/helm/charts/onyx/values-lite.yaml",
+    "description": "Lite deployment Helm values and service assumptions"
+  },
+  {
+    "scope": [],
+    "path": "deployment/docker_compose/docker-compose.onyx-lite.yml",
+    "description": "Lite deployment Docker Compose overlay and disabled service behavior"
+  }
+]

.greptile/rules.md

@@ -0,0 +1,39 @@
+# Greptile Review Rules
+
+## Type Annotations
+
+Use explicit type annotations for variables to enhance code clarity, especially when moving type hints around in the code.
+
+## Best Practices
+
+Use `contributing_guides/best_practices.md` as core review context. Prefer consistency with existing patterns, fix issues in code you touch, avoid tacking new features onto muddy interfaces, fail loudly instead of silently swallowing errors, keep code strictly typed, preserve clear state boundaries, remove duplicate or dead logic, break up overly long functions, avoid hidden import-time side effects, respect module boundaries, and favor correctness-by-construction over relying on callers to use an API correctly.
+
+## TODOs
+
+Whenever a TODO is added, there must always be an associated name or ticket with that TODO in the style of `TODO(name): ...` or `TODO(1234): ...`
+
+## Debugging Code
+
+Remove temporary debugging code before merging to production, especially tenant-specific debugging logs.
+
+## Hardcoded Booleans
+
+When hardcoding a boolean variable to a constant value, remove the variable entirely and clean up all places where it's used rather than just setting it to a constant.
+
+## Multi-tenant vs Single-tenant
+
+Code changes must consider both multi-tenant and single-tenant deployments. In multi-tenant mode, preserve tenant isolation, ensure tenant context is propagated correctly, and avoid assumptions that only hold for a single shared schema or globally shared state. In single-tenant mode, avoid introducing unnecessary tenant-specific requirements or cloud-only control-plane dependencies.
+
+## Nginx Routing — New Backend Routes
+
+Whenever a new backend route is added that does NOT start with `/api`, it must also be explicitly added to ALL nginx configs:
+- `deployment/helm/charts/onyx/templates/nginx-conf.yaml` (Helm/k8s)
+- `deployment/data/nginx/app.conf.template` (docker-compose dev)
+- `deployment/data/nginx/app.conf.template.prod` (docker-compose prod)
+- `deployment/data/nginx/app.conf.template.no-letsencrypt` (docker-compose no-letsencrypt)
+
+Routes not starting with `/api` are not caught by the existing `^/(api|openapi\.json)` location block and will fall through to `location /`, which proxies to the Next.js web server and returns an HTML 404. The new location block must be placed before the `/api` block. Examples of routes that need this treatment: `/scim`, `/mcp`.
+
+## Full vs Lite Deployments
+
+Code changes must consider both regular Onyx deployments and Onyx lite deployments. Lite deployments disable the vector DB, Redis, model servers, and background workers by default, use PostgreSQL-backed cache/auth/file storage, and rely on the API server to handle background work. Do not assume those services are available unless the code path is explicitly limited to full deployments.

.pre-commit-config.yaml

@@ -122,7 +122,7 @@ repos:
     rev: 5d1e709b7be35cb2025444e19de266b056b7b7ee # frozen: v2.10.1
     hooks:
       - id: golangci-lint
-        language_version: "1.26.0"
+        language_version: "1.26.1"
         entry: bash -c "find . -name go.mod -not -path './.venv/*' -print0 | xargs -0 -I{} bash -c 'cd \"$(dirname {})\" && golangci-lint run ./...'"
 
   - repo: https://github.com/astral-sh/ruff-pre-commit

.vscode/launch.json

@@ -117,7 +117,8 @@
       "presentation": {
         "group": "2"
       },
-      "consoleTitle": "API Server Console"
+      "consoleTitle": "API Server Console",
+      "justMyCode": false
     },
     {
       "name": "Slack Bot",
@@ -268,7 +269,8 @@
       "presentation": {
         "group": "2"
       },
-      "consoleTitle": "Celery heavy Console"
+      "consoleTitle": "Celery heavy Console",
+      "justMyCode": false
     },
     {
       "name": "Celery kg_processing",
@@ -355,7 +357,8 @@
       "presentation": {
         "group": "2"
       },
-      "consoleTitle": "Celery user_file_processing Console"
+      "consoleTitle": "Celery user_file_processing Console",
+      "justMyCode": false
     },
     {
       "name": "Celery docfetching",
@@ -413,7 +416,8 @@
       "presentation": {
         "group": "2"
       },
-      "consoleTitle": "Celery docprocessing Console"
+      "consoleTitle": "Celery docprocessing Console",
+      "justMyCode": false
     },
     {
       "name": "Celery beat",

AGENTS.md

@@ -167,284 +167,7 @@ web/
 
 ## Frontend Standards
 
-### 1. Import Standards
-
-**Always use absolute imports with the `@` prefix.**
-
-**Reason:** Moving files around becomes easier since you don't also have to update those import statements. This makes modifications to the codebase much nicer.
-
-```typescript
-// ✅ Good
-import { Button } from "@/components/ui/button";
-import { useAuth } from "@/hooks/useAuth";
-import { Text } from "@/refresh-components/texts/Text";
-
-// ❌ Bad
-import { Button } from "../../../components/ui/button";
-import { useAuth } from "./hooks/useAuth";
-```
-
-### 2. React Component Functions
-
-**Prefer regular functions over arrow functions for React components.**
-
-**Reason:** Functions just become easier to read.
-
-```typescript
-// ✅ Good
-function UserProfile({ userId }: UserProfileProps) {
-  return <div>User Profile</div>
-}
-
-// ❌ Bad
-const UserProfile = ({ userId }: UserProfileProps) => {
-  return <div>User Profile</div>
-}
-```
-
-### 3. Props Interface Extraction
-
-**Extract prop types into their own interface definitions.**
-
-**Reason:** Functions just become easier to read.
-
-```typescript
-// ✅ Good
-interface UserCardProps {
-  user: User
-  showActions?: boolean
-  onEdit?: (userId: string) => void
-}
-
-function UserCard({ user, showActions = false, onEdit }: UserCardProps) {
-  return <div>User Card</div>
-}
-
-// ❌ Bad
-function UserCard({
-  user,
-  showActions = false,
-  onEdit
-}: {
-  user: User
-  showActions?: boolean
-  onEdit?: (userId: string) => void
-}) {
-  return <div>User Card</div>
-}
-```
-
-### 4. Spacing Guidelines
-
-**Prefer padding over margins for spacing.**
-
-**Reason:** We want to consolidate usage to paddings instead of margins.
-
-```typescript
-// ✅ Good
-<div className="p-4 space-y-2">
-  <div className="p-2">Content</div>
-</div>
-
-// ❌ Bad
-<div className="m-4 space-y-2">
-  <div className="m-2">Content</div>
-</div>
-```
-
-### 5. Tailwind Dark Mode
-
-**Strictly forbid using the `dark:` modifier in Tailwind classes, except for logo icon handling.**
-
-**Reason:** The `colors.css` file already, VERY CAREFULLY, defines what the exact opposite colour of each light-mode colour is. Overriding this behaviour is VERY bad and will lead to horrible UI breakages.
-
-**Exception:** The `createLogoIcon` helper in `web/src/components/icons/icons.tsx` uses `dark:` modifiers (`dark:invert`, `dark:hidden`, `dark:block`) to handle third-party logo icons that cannot automatically adapt through `colors.css`. This is the ONLY acceptable use of dark mode modifiers.
-
-```typescript
-// ✅ Good - Standard components use `tailwind-themes/tailwind.config.js` / `src/app/css/colors.css`
-<div className="bg-background-neutral-03 text-text-02">
-  Content
-</div>
-
-// ✅ Good - Logo icons with dark mode handling via createLogoIcon
-export const GithubIcon = createLogoIcon(githubLightIcon, {
-  monochromatic: true,  // Will apply dark:invert internally
-});
-
-export const GitbookIcon = createLogoIcon(gitbookLightIcon, {
-  darkSrc: gitbookDarkIcon,  // Will use dark:hidden/dark:block internally
-});
-
-// ❌ Bad - Manual dark mode overrides
-<div className="bg-white dark:bg-black text-black dark:text-white">
-  Content
-</div>
-```
-
-### 6. Class Name Utilities
-
-**Use the `cn` utility instead of raw string formatting for classNames.**
-
-**Reason:** `cn`s are easier to read. They also allow for more complex types (i.e., string-arrays) to get formatted properly (it flattens each element in that string array down). As a result, it can allow things such as conditionals (i.e., `myCondition && "some-tailwind-class"`, which evaluates to `false` when `myCondition` is `false`) to get filtered out.
-
-```typescript
-import { cn } from '@/lib/utils'
-
-// ✅ Good
-<div className={cn(
-  'base-class',
-  isActive && 'active-class',
-  className
-)}>
-  Content
-</div>
-
-// ❌ Bad
-<div className={`base-class ${isActive ? 'active-class' : ''} ${className}`}>
-  Content
-</div>
-```
-
-### 7. Custom Hooks Organization
-
-**Follow a "hook-per-file" layout. Each hook should live in its own file within `web/src/hooks`.**
-
-**Reason:** This is just a layout preference. Keeps code clean.
-
-```typescript
-// web/src/hooks/useUserData.ts
-export function useUserData(userId: string) {
-  // hook implementation
-}
-
-// web/src/hooks/useLocalStorage.ts
-export function useLocalStorage<T>(key: string, initialValue: T) {
-  // hook implementation
-}
-```
-
-### 8. Icon Usage
-
-**ONLY use icons from the `web/src/icons` directory. Do NOT use icons from `react-icons`, `lucide`, or other external libraries.**
-
-**Reason:** We have a very carefully curated selection of icons that match our Onyx guidelines. We do NOT want to muddy those up with different aesthetic stylings.
-
-```typescript
-// ✅ Good
-import SvgX from "@/icons/x";
-import SvgMoreHorizontal from "@/icons/more-horizontal";
-
-// ❌ Bad
-import { User } from "lucide-react";
-import { FiSearch } from "react-icons/fi";
-```
-
-**Missing Icons**: If an icon is needed but doesn't exist in the `web/src/icons` directory, import it from Figma using the Figma MCP tool and add it to the icons directory.
-If you need help with this step, reach out to `raunak@onyx.app`.
-
-### 9. Text Rendering
-
-**Prefer using the `refresh-components/texts/Text` component for all text rendering. Avoid "naked" text nodes.**
-
-**Reason:** The `Text` component is fully compliant with the stylings provided in Figma. It provides easy utilities to specify the text-colour and font-size in the form of flags. Super duper easy.
-
-```typescript
-// ✅ Good
-import { Text } from '@/refresh-components/texts/Text'
-
-function UserCard({ name }: { name: string }) {
-  return (
-    <Text
-      {/* The `text03` flag makes the text it renders to be coloured the 3rd-scale grey */}
-      text03
-      {/* The `mainAction` flag makes the text it renders to be "main-action" font + line-height + weightage, as described in the Figma */}
-      mainAction
-    >
-      {name}
-    </Text>
-  )
-}
-
-// ❌ Bad
-function UserCard({ name }: { name: string }) {
-  return (
-    <div>
-      <h2>{name}</h2>
-      <p>User details</p>
-    </div>
-  )
-}
-```
-
-### 10. Component Usage
-
-**Heavily avoid raw HTML input components. Always use components from the `web/src/refresh-components` or `web/lib/opal/src` directory.**
-
-**Reason:** We've put in a lot of effort to unify the components that are rendered in the Onyx app. Using raw components breaks the entire UI of the application, and leaves it in a muddier state than before.
-
-```typescript
-// ✅ Good
-import Button from '@/refresh-components/buttons/Button'
-import InputTypeIn from '@/refresh-components/inputs/InputTypeIn'
-import SvgPlusCircle from '@/icons/plus-circle'
-
-function ContactForm() {
-  return (
-    <form>
-      <InputTypeIn placeholder="Search..." />
-      <Button type="submit" leftIcon={SvgPlusCircle}>Submit</Button>
-    </form>
-  )
-}
-
-// ❌ Bad
-function ContactForm() {
-  return (
-    <form>
-      <input placeholder="Name" />
-      <textarea placeholder="Message" />
-      <button type="submit">Submit</button>
-    </form>
-  )
-}
-```
-
-### 11. Colors
-
-**Always use custom overrides for colors and borders rather than built in Tailwind CSS colors. These overrides live in `web/tailwind-themes/tailwind.config.js`.**
-
-**Reason:** Our custom color system uses CSS variables that automatically handle dark mode and maintain design consistency across the app. Standard Tailwind colors bypass this system.
-
-**Available color categories:**
-
-- **Text:** `text-01` through `text-05`, `text-inverted-XX`
-- **Backgrounds:** `background-neutral-XX`, `background-tint-XX` (and inverted variants)
-- **Borders:** `border-01` through `border-05`, `border-inverted-XX`
-- **Actions:** `action-link-XX`, `action-danger-XX`
-- **Status:** `status-info-XX`, `status-success-XX`, `status-warning-XX`, `status-error-XX`
-- **Theme:** `theme-primary-XX`, `theme-red-XX`, `theme-blue-XX`, etc.
-
-```typescript
-// ✅ Good - Use custom Onyx color classes
-<div className="bg-background-neutral-01 border border-border-02" />
-<div className="bg-background-tint-02 border border-border-01" />
-<div className="bg-status-success-01" />
-<div className="bg-action-link-01" />
-<div className="bg-theme-primary-05" />
-
-// ❌ Bad - Do NOT use standard Tailwind colors
-<div className="bg-gray-100 border border-gray-300 text-gray-600" />
-<div className="bg-white border border-slate-200" />
-<div className="bg-green-100 text-green-700" />
-<div className="bg-blue-100 text-blue-600" />
-<div className="bg-indigo-500" />
-```
-
-### 12. Data Fetching
-
-**Prefer using `useSWR` for data fetching. Data should generally be fetched on the client side. Components that need data should display a loader / placeholder while waiting for that data. Prefer loading data within the component that needs it rather than at the top level and passing it down.**
-
-**Reason:** Client side fetching allows us to load the skeleton of the page without waiting for data to load, leading to a snappier UX. Loading data where needed reduces dependencies between a component and its parent component(s).
+Frontend standards for the `web/` and `desktop/` projects live in `web/AGENTS.md`.
 
 ## Database & Migrations
 

README.md

@@ -35,7 +35,7 @@ Onyx comes loaded with advanced features like Agents, Web Search, RAG, MCP, Deep
 > [!TIP]
 > Run Onyx with one command (or see deployment section below):
 > ```
-> curl -fsSL https://raw.githubusercontent.com/onyx-dot-app/onyx/main/deployment/docker_compose/install.sh > install.sh && chmod +x install.sh && ./install.sh
+> curl -fsSL https://onyx.app/install_onyx.sh | bash
 > ```
 
 ****

backend/Dockerfile

@@ -47,6 +47,8 @@ RUN apt-get update && \
         gcc \
         nano \
         vim \
+        # Install procps so kubernetes exec sessions can use ps aux for debugging
+        procps \
         libjemalloc2 \
         && \
     rm -rf /var/lib/apt/lists/* && \

backend/alembic/versions/1d78c0ca7853_remove_voice_provider_deleted_column.py

@@ -0,0 +1,35 @@
+"""remove voice_provider deleted column
+
+Revision ID: 1d78c0ca7853
+Revises: a3f8b2c1d4e5
+Create Date: 2026-03-26 11:30:53.883127
+
+"""
+
+from alembic import op
+import sqlalchemy as sa
+
+
+# revision identifiers, used by Alembic.
+revision = "1d78c0ca7853"
+down_revision = "a3f8b2c1d4e5"
+branch_labels = None
+depends_on = None
+
+
+def upgrade() -> None:
+    # Hard-delete any soft-deleted rows before dropping the column
+    op.execute("DELETE FROM voice_provider WHERE deleted = true")
+    op.drop_column("voice_provider", "deleted")
+
+
+def downgrade() -> None:
+    op.add_column(
+        "voice_provider",
+        sa.Column(
+            "deleted",
+            sa.Boolean(),
+            nullable=False,
+            server_default=sa.text("false"),
+        ),
+    )

backend/alembic/versions/25a5501dc766_group_permissions_phase1.py

@@ -0,0 +1,109 @@
+"""group_permissions_phase1
+
+Revision ID: 25a5501dc766
+Revises: b728689f45b1
+Create Date: 2026-03-23 11:41:25.557442
+
+"""
+
+from alembic import op
+import fastapi_users_db_sqlalchemy
+import sqlalchemy as sa
+
+from onyx.db.enums import AccountType
+from onyx.db.enums import GrantSource
+from onyx.db.enums import Permission
+
+
+# revision identifiers, used by Alembic.
+revision = "25a5501dc766"
+down_revision = "b728689f45b1"
+branch_labels = None
+depends_on = None
+
+
+def upgrade() -> None:
+    # 1. Add account_type column to user table (nullable for now).
+    #    TODO(subash): backfill account_type for existing rows and add NOT NULL.
+    op.add_column(
+        "user",
+        sa.Column(
+            "account_type",
+            sa.Enum(AccountType, native_enum=False),
+            nullable=True,
+        ),
+    )
+
+    # 2. Add is_default column to user_group table
+    op.add_column(
+        "user_group",
+        sa.Column(
+            "is_default",
+            sa.Boolean(),
+            nullable=False,
+            server_default=sa.false(),
+        ),
+    )
+
+    # 3. Create permission_grant table
+    op.create_table(
+        "permission_grant",
+        sa.Column("id", sa.Integer(), autoincrement=True, nullable=False),
+        sa.Column("group_id", sa.Integer(), nullable=False),
+        sa.Column(
+            "permission",
+            sa.Enum(Permission, native_enum=False),
+            nullable=False,
+        ),
+        sa.Column(
+            "grant_source",
+            sa.Enum(GrantSource, native_enum=False),
+            nullable=False,
+        ),
+        sa.Column(
+            "granted_by",
+            fastapi_users_db_sqlalchemy.generics.GUID(),
+            nullable=True,
+        ),
+        sa.Column(
+            "granted_at",
+            sa.DateTime(timezone=True),
+            server_default=sa.func.now(),
+            nullable=False,
+        ),
+        sa.Column(
+            "is_deleted",
+            sa.Boolean(),
+            nullable=False,
+            server_default=sa.false(),
+        ),
+        sa.PrimaryKeyConstraint("id"),
+        sa.ForeignKeyConstraint(
+            ["group_id"],
+            ["user_group.id"],
+            ondelete="CASCADE",
+        ),
+        sa.ForeignKeyConstraint(
+            ["granted_by"],
+            ["user.id"],
+            ondelete="SET NULL",
+        ),
+        sa.UniqueConstraint(
+            "group_id", "permission", name="uq_permission_grant_group_permission"
+        ),
+    )
+
+    # 4. Index on user__user_group(user_id) — existing composite PK
+    #    has user_group_id as leading column; user-filtered queries need this
+    op.create_index(
+        "ix_user__user_group_user_id",
+        "user__user_group",
+        ["user_id"],
+    )
+
+
+def downgrade() -> None:
+    op.drop_index("ix_user__user_group_user_id", table_name="user__user_group")
+    op.drop_table("permission_grant")
+    op.drop_column("user_group", "is_default")
+    op.drop_column("user", "account_type")

backend/alembic/versions/689433b0d8de_add_hook_and_hook_execution_log_tables.py

@@ -0,0 +1,103 @@
+"""add_hook_and_hook_execution_log_tables
+
+Revision ID: 689433b0d8de
+Revises: 93a2e195e25c
+Create Date: 2026-03-13 11:25:06.547474
+
+"""
+
+from alembic import op
+import sqlalchemy as sa
+from sqlalchemy.dialects.postgresql import UUID as PGUUID
+
+
+# revision identifiers, used by Alembic.
+revision = "689433b0d8de"
+down_revision = "93a2e195e25c"
+branch_labels = None
+depends_on = None
+
+
+def upgrade() -> None:
+    op.create_table(
+        "hook",
+        sa.Column("id", sa.Integer(), nullable=False),
+        sa.Column("name", sa.String(), nullable=False),
+        sa.Column(
+            "hook_point",
+            sa.Enum("document_ingestion", "query_processing", native_enum=False),
+            nullable=False,
+        ),
+        sa.Column("endpoint_url", sa.Text(), nullable=True),
+        sa.Column("api_key", sa.LargeBinary(), nullable=True),
+        sa.Column("is_reachable", sa.Boolean(), nullable=True),
+        sa.Column(
+            "fail_strategy",
+            sa.Enum("hard", "soft", native_enum=False),
+            nullable=False,
+        ),
+        sa.Column("timeout_seconds", sa.Float(), nullable=False),
+        sa.Column(
+            "is_active", sa.Boolean(), nullable=False, server_default=sa.text("false")
+        ),
+        sa.Column(
+            "deleted", sa.Boolean(), nullable=False, server_default=sa.text("false")
+        ),
+        sa.Column("creator_id", PGUUID(as_uuid=True), nullable=True),
+        sa.Column(
+            "created_at",
+            sa.DateTime(timezone=True),
+            server_default=sa.text("now()"),
+            nullable=False,
+        ),
+        sa.Column(
+            "updated_at",
+            sa.DateTime(timezone=True),
+            server_default=sa.text("now()"),
+            nullable=False,
+        ),
+        sa.ForeignKeyConstraint(["creator_id"], ["user.id"], ondelete="SET NULL"),
+        sa.PrimaryKeyConstraint("id"),
+    )
+    op.create_index(
+        "ix_hook_one_non_deleted_per_point",
+        "hook",
+        ["hook_point"],
+        unique=True,
+        postgresql_where=sa.text("deleted = false"),
+    )
+
+    op.create_table(
+        "hook_execution_log",
+        sa.Column("id", sa.Integer(), nullable=False),
+        sa.Column("hook_id", sa.Integer(), nullable=False),
+        sa.Column(
+            "is_success",
+            sa.Boolean(),
+            nullable=False,
+        ),
+        sa.Column("error_message", sa.Text(), nullable=True),
+        sa.Column("status_code", sa.Integer(), nullable=True),
+        sa.Column("duration_ms", sa.Integer(), nullable=True),
+        sa.Column(
+            "created_at",
+            sa.DateTime(timezone=True),
+            server_default=sa.text("now()"),
+            nullable=False,
+        ),
+        sa.ForeignKeyConstraint(["hook_id"], ["hook.id"], ondelete="CASCADE"),
+        sa.PrimaryKeyConstraint("id"),
+    )
+    op.create_index("ix_hook_execution_log_hook_id", "hook_execution_log", ["hook_id"])
+    op.create_index(
+        "ix_hook_execution_log_created_at", "hook_execution_log", ["created_at"]
+    )
+
+
+def downgrade() -> None:
+    op.drop_index("ix_hook_execution_log_created_at", table_name="hook_execution_log")
+    op.drop_index("ix_hook_execution_log_hook_id", table_name="hook_execution_log")
+    op.drop_table("hook_execution_log")
+
+    op.drop_index("ix_hook_one_non_deleted_per_point", table_name="hook")
+    op.drop_table("hook")

backend/alembic/versions/8188861f4e92_csv_to_tabular_chat_file_type.py

@@ -0,0 +1,54 @@
+"""csv to tabular chat file type
+
+Revision ID: 8188861f4e92
+Revises: d8cdfee5df80
+Create Date: 2026-03-31 19:23:05.753184
+
+"""
+
+from alembic import op
+
+
+# revision identifiers, used by Alembic.
+revision = "8188861f4e92"
+down_revision = "d8cdfee5df80"
+branch_labels = None
+depends_on = None
+
+
+def upgrade() -> None:
+    op.execute(
+        """
+        UPDATE chat_message
+        SET files = (
+            SELECT jsonb_agg(
+                CASE
+                    WHEN elem->>'type' = 'csv'
+                    THEN jsonb_set(elem, '{type}', '"tabular"')
+                    ELSE elem
+                END
+            )
+            FROM jsonb_array_elements(files) AS elem
+        )
+        WHERE files::text LIKE '%"type": "csv"%'
+        """
+    )
+
+
+def downgrade() -> None:
+    op.execute(
+        """
+        UPDATE chat_message
+        SET files = (
+            SELECT jsonb_agg(
+                CASE
+                    WHEN elem->>'type' = 'tabular'
+                    THEN jsonb_set(elem, '{type}', '"csv"')
+                    ELSE elem
+                END
+            )
+            FROM jsonb_array_elements(files) AS elem
+        )
+        WHERE files::text LIKE '%"type": "tabular"%'
+        """
+    )

backend/alembic/versions/a3f8b2c1d4e5_add_preferred_response_id_to_chat_message.py

@@ -0,0 +1,36 @@
+"""add preferred_response_id and model_display_name to chat_message
+
+Revision ID: a3f8b2c1d4e5
+Create Date: 2026-03-22
+
+"""
+
+from alembic import op
+import sqlalchemy as sa
+
+# revision identifiers, used by Alembic.
+revision = "a3f8b2c1d4e5"
+down_revision = "25a5501dc766"
+branch_labels = None
+depends_on = None
+
+
+def upgrade() -> None:
+    op.add_column(
+        "chat_message",
+        sa.Column(
+            "preferred_response_id",
+            sa.Integer(),
+            sa.ForeignKey("chat_message.id", ondelete="SET NULL"),
+            nullable=True,
+        ),
+    )
+    op.add_column(
+        "chat_message",
+        sa.Column("model_display_name", sa.String(), nullable=True),
+    )
+
+
+def downgrade() -> None:
+    op.drop_column("chat_message", "model_display_name")
+    op.drop_column("chat_message", "preferred_response_id")

backend/alembic/versions/b728689f45b1_rename_persona_is_visible_to_is_listed_.py

@@ -0,0 +1,26 @@
+"""rename persona is_visible to is_listed and featured to is_featured
+
+Revision ID: b728689f45b1
+Revises: 689433b0d8de
+Create Date: 2026-03-23 12:36:26.607305
+
+"""
+
+from alembic import op
+
+
+# revision identifiers, used by Alembic.
+revision = "b728689f45b1"
+down_revision = "689433b0d8de"
+branch_labels = None
+depends_on = None
+
+
+def upgrade() -> None:
+    op.alter_column("persona", "is_visible", new_column_name="is_listed")
+    op.alter_column("persona", "featured", new_column_name="is_featured")
+
+
+def downgrade() -> None:
+    op.alter_column("persona", "is_listed", new_column_name="is_visible")
+    op.alter_column("persona", "is_featured", new_column_name="featured")

backend/alembic/versions/d8cdfee5df80_add_skipped_to_userfilestatus.py

@@ -0,0 +1,55 @@
+"""add skipped to userfilestatus
+
+Revision ID: d8cdfee5df80
+Revises: 1d78c0ca7853
+Create Date: 2026-04-01 10:47:12.593950
+
+"""
+
+from alembic import op
+import sqlalchemy as sa
+
+
+# revision identifiers, used by Alembic.
+revision = "d8cdfee5df80"
+down_revision = "1d78c0ca7853"
+branch_labels = None
+depends_on = None
+
+
+TABLE = "user_file"
+COLUMN = "status"
+CONSTRAINT_NAME = "ck_user_file_status"
+
+OLD_VALUES = ("PROCESSING", "INDEXING", "COMPLETED", "FAILED", "CANCELED", "DELETING")
+NEW_VALUES = (
+    "PROCESSING",
+    "INDEXING",
+    "COMPLETED",
+    "SKIPPED",
+    "FAILED",
+    "CANCELED",
+    "DELETING",
+)
+
+
+def _drop_status_check_constraint() -> None:
+    inspector = sa.inspect(op.get_bind())
+    for constraint in inspector.get_check_constraints(TABLE):
+        if COLUMN in constraint.get("sqltext", ""):
+            constraint_name = constraint["name"]
+            if constraint_name is not None:
+                op.drop_constraint(constraint_name, TABLE, type_="check")
+
+
+def upgrade() -> None:
+    _drop_status_check_constraint()
+    in_clause = ", ".join(f"'{v}'" for v in NEW_VALUES)
+    op.create_check_constraint(CONSTRAINT_NAME, TABLE, f"{COLUMN} IN ({in_clause})")
+
+
+def downgrade() -> None:
+    op.execute(f"UPDATE {TABLE} SET {COLUMN} = 'COMPLETED' WHERE {COLUMN} = 'SKIPPED'")
+    _drop_status_check_constraint()
+    in_clause = ", ".join(f"'{v}'" for v in OLD_VALUES)
+    op.create_check_constraint(CONSTRAINT_NAME, TABLE, f"{COLUMN} IN ({in_clause})")

backend/alembic/versions/e7f8a9b0c1d2_create_anonymous_user.py

@@ -36,6 +36,56 @@ TABLES_WITH_USER_ID = [
 ]
 
 
+def _dedupe_null_notifications(connection: sa.Connection) -> None:
+    # Multiple NULL-owned notifications can exist because the unique index treats
+    # NULL user_id values as distinct. Before migrating them to the anonymous
+    # user, collapse duplicates and remove rows that would conflict with an
+    # already-existing anonymous notification.
+    result = connection.execute(
+        sa.text(
+            """
+            WITH ranked_null_notifications AS (
+                SELECT
+                    id,
+                    ROW_NUMBER() OVER (
+                        PARTITION BY notif_type, COALESCE(additional_data, '{}'::jsonb)
+                        ORDER BY first_shown DESC, last_shown DESC, id DESC
+                    ) AS row_num
+                FROM notification
+                WHERE user_id IS NULL
+            )
+            DELETE FROM notification
+            WHERE id IN (
+                SELECT id
+                FROM ranked_null_notifications
+                WHERE row_num > 1
+            )
+            """
+        )
+    )
+    if result.rowcount > 0:
+        print(f"Deleted {result.rowcount} duplicate NULL-owned notifications")
+
+    result = connection.execute(
+        sa.text(
+            """
+            DELETE FROM notification AS null_owned
+            USING notification AS anonymous_owned
+            WHERE null_owned.user_id IS NULL
+              AND anonymous_owned.user_id = :user_id
+              AND null_owned.notif_type = anonymous_owned.notif_type
+              AND COALESCE(null_owned.additional_data, '{}'::jsonb) =
+                  COALESCE(anonymous_owned.additional_data, '{}'::jsonb)
+            """
+        ),
+        {"user_id": ANONYMOUS_USER_UUID},
+    )
+    if result.rowcount > 0:
+        print(
+            f"Deleted {result.rowcount} NULL-owned notifications that conflict with existing anonymous-owned notifications"
+        )
+
+
 def upgrade() -> None:
     """
     Create the anonymous user for anonymous access feature.
@@ -65,7 +115,12 @@ def upgrade() -> None:
 
     # Migrate any remaining user_id=NULL records to anonymous user
     for table in TABLES_WITH_USER_ID:
-        try:
+        # Dedup notifications outside the savepoint so deletions persist
+        # even if the subsequent UPDATE rolls back
+        if table == "notification":
+            _dedupe_null_notifications(connection)
+
+        with connection.begin_nested():
             # Exclude public credential (id=0) which must remain user_id=NULL
             # Exclude builtin tools (in_code_tool_id IS NOT NULL) which must remain user_id=NULL
             # Exclude builtin personas (builtin_persona=True) which must remain user_id=NULL
@@ -80,6 +135,7 @@ def upgrade() -> None:
                 condition = "user_id IS NULL AND is_public = false"
             else:
                 condition = "user_id IS NULL"
+
             result = connection.execute(
                 sa.text(
                     f"""
@@ -92,19 +148,19 @@ def upgrade() -> None:
             )
             if result.rowcount > 0:
                 print(f"Updated {result.rowcount} rows in {table} to anonymous user")
-        except Exception as e:
-            print(f"Skipping {table}: {e}")
 
 
 def downgrade() -> None:
     """
     Set anonymous user's records back to NULL and delete the anonymous user.
+
+    Note: Duplicate NULL-owned notifications removed during upgrade are not restored.
     """
     connection = op.get_bind()
 
     # Set records back to NULL
     for table in TABLES_WITH_USER_ID:
-        try:
+        with connection.begin_nested():
             connection.execute(
                 sa.text(
                     f"""
@@ -115,8 +171,6 @@ def downgrade() -> None:
                 ),
                 {"user_id": ANONYMOUS_USER_UUID},
             )
-        except Exception:
-            pass
 
     # Delete the anonymous user
     connection.execute(

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

@@ -5,6 +5,7 @@ from onyx.background.celery.apps.primary import celery_app
 celery_app.autodiscover_tasks(
     app_base.filter_task_modules(
         [
+            "ee.onyx.background.celery.tasks.hooks",
             "ee.onyx.background.celery.tasks.doc_permission_syncing",
             "ee.onyx.background.celery.tasks.external_group_syncing",
             "ee.onyx.background.celery.tasks.cloud",

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

@@ -55,6 +55,15 @@ ee_tasks_to_schedule: list[dict] = []
 
 if not MULTI_TENANT:
     ee_tasks_to_schedule = [
+        {
+            "name": "hook-execution-log-cleanup",
+            "task": OnyxCeleryTask.HOOK_EXECUTION_LOG_CLEANUP_TASK,
+            "schedule": timedelta(days=1),
+            "options": {
+                "priority": OnyxCeleryPriority.LOW,
+                "expires": BEAT_EXPIRES_DEFAULT,
+            },
+        },
         {
             "name": "autogenerate-usage-report",
             "task": OnyxCeleryTask.GENERATE_USAGE_REPORT_TASK,

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

@@ -28,6 +28,7 @@ from onyx.access.models import DocExternalAccess
 from onyx.access.models import ElementExternalAccess
 from onyx.background.celery.apps.app_base import task_logger
 from onyx.background.celery.celery_redis import celery_find_task
+from onyx.background.celery.celery_redis import celery_get_broker_client
 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
@@ -187,7 +188,6 @@ def check_for_doc_permissions_sync(self: Task, *, tenant_id: str) -> bool | None
     # (which lives on a different db number)
     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_DOC_PERMISSIONS_SYNC_BEAT_LOCK,
@@ -227,6 +227,7 @@ def check_for_doc_permissions_sync(self: Task, *, tenant_id: str) -> bool | None
             # tasks can be in the queue in redis, in reserved tasks (prefetched by the worker),
             # or be currently executing
             try:
+                r_celery = celery_get_broker_client(self.app)
                 validate_permission_sync_fences(
                     tenant_id, r, r_replica, r_celery, lock_beat
                 )
@@ -473,6 +474,8 @@ def connector_permission_sync_generator_task(
             cc_pair = get_connector_credential_pair_from_id(
                 db_session=db_session,
                 cc_pair_id=cc_pair_id,
+                eager_load_connector=True,
+                eager_load_credential=True,
             )
             if cc_pair is None:
                 raise ValueError(

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

@@ -29,6 +29,7 @@ from ee.onyx.external_permissions.sync_params import (
 from ee.onyx.external_permissions.sync_params import get_source_perm_sync_config
 from onyx.background.celery.apps.app_base import task_logger
 from onyx.background.celery.celery_redis import celery_find_task
+from onyx.background.celery.celery_redis import celery_get_broker_client
 from onyx.background.celery.celery_redis import celery_get_unacked_task_ids
 from onyx.background.celery.tasks.beat_schedule import CLOUD_BEAT_MULTIPLIER_DEFAULT
 from onyx.background.error_logging import emit_background_error
@@ -162,7 +163,6 @@ def check_for_external_group_sync(self: Task, *, tenant_id: str) -> bool | None:
     # (which lives on a different db number)
     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_EXTERNAL_GROUP_SYNC_BEAT_LOCK,
@@ -221,6 +221,7 @@ def check_for_external_group_sync(self: Task, *, tenant_id: str) -> bool | None:
             # tasks can be in the queue in redis, in reserved tasks (prefetched by the worker),
             # or be currently executing
             try:
+                r_celery = celery_get_broker_client(self.app)
                 validate_external_group_sync_fences(
                     tenant_id, self.app, r, r_replica, r_celery, lock_beat
                 )

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

@@ -0,0 +1,35 @@
+from celery import shared_task
+
+from onyx.configs.app_configs import JOB_TIMEOUT
+from onyx.configs.constants import OnyxCeleryTask
+from onyx.db.engine.sql_engine import get_session_with_current_tenant
+from onyx.db.hook import cleanup_old_execution_logs__no_commit
+from onyx.utils.logger import setup_logger
+
+logger = setup_logger()
+
+_HOOK_EXECUTION_LOG_RETENTION_DAYS: int = 30
+
+
+@shared_task(
+    name=OnyxCeleryTask.HOOK_EXECUTION_LOG_CLEANUP_TASK,
+    ignore_result=True,
+    soft_time_limit=JOB_TIMEOUT,
+    trail=False,
+)
+def hook_execution_log_cleanup_task(*, tenant_id: str) -> None:  # noqa: ARG001
+    try:
+        with get_session_with_current_tenant() as db_session:
+            deleted: int = cleanup_old_execution_logs__no_commit(
+                db_session=db_session,
+                max_age_days=_HOOK_EXECUTION_LOG_RETENTION_DAYS,
+            )
+            db_session.commit()
+            if deleted:
+                logger.info(
+                    f"Deleted {deleted} hook execution log(s) older than "
+                    f"{_HOOK_EXECUTION_LOG_RETENTION_DAYS} days."
+                )
+    except Exception:
+        logger.exception("Failed to clean up hook execution logs")
+        raise

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

@@ -13,6 +13,7 @@ from redis.lock import Lock as RedisLock
 from ee.onyx.server.tenants.provisioning import setup_tenant
 from ee.onyx.server.tenants.schema_management import create_schema_if_not_exists
 from ee.onyx.server.tenants.schema_management import get_current_alembic_version
+from ee.onyx.server.tenants.schema_management import run_alembic_migrations
 from onyx.background.celery.apps.app_base import task_logger
 from onyx.configs.app_configs import TARGET_AVAILABLE_TENANTS
 from onyx.configs.constants import ONYX_CLOUD_TENANT_ID
@@ -25,13 +26,14 @@ from onyx.redis.redis_pool import get_redis_client
 from shared_configs.configs import MULTI_TENANT
 from shared_configs.configs import TENANT_ID_PREFIX
 
-# Default number of pre-provisioned tenants to maintain
-DEFAULT_TARGET_AVAILABLE_TENANTS = 5
+# Maximum tenants to provision in a single task run.
+# Each tenant takes ~80s (alembic migrations), so 5 tenants ≈ 7 minutes.
+_MAX_TENANTS_PER_RUN = 5
 
-# Soft time limit for tenant pre-provisioning tasks (in seconds)
-_TENANT_PROVISIONING_SOFT_TIME_LIMIT = 60 * 5  # 5 minutes
-# Hard time limit for tenant pre-provisioning tasks (in seconds)
-_TENANT_PROVISIONING_TIME_LIMIT = 60 * 10  # 10 minutes
+# Time limits sized for worst-case: provisioning up to _MAX_TENANTS_PER_RUN new tenants
+# (~90s each) plus migrating up to TARGET_AVAILABLE_TENANTS pool tenants (~90s each).
+_TENANT_PROVISIONING_SOFT_TIME_LIMIT = 60 * 20  # 20 minutes
+_TENANT_PROVISIONING_TIME_LIMIT = 60 * 25  # 25 minutes
 
 
 @shared_task(
@@ -58,7 +60,7 @@ def check_available_tenants(self: Task) -> None:  # noqa: ARG001
     r = get_redis_client(tenant_id=ONYX_CLOUD_TENANT_ID)
     lock_check: RedisLock = r.lock(
         OnyxRedisLocks.CHECK_AVAILABLE_TENANTS_LOCK,
-        timeout=_TENANT_PROVISIONING_SOFT_TIME_LIMIT,
+        timeout=_TENANT_PROVISIONING_TIME_LIMIT,
     )
 
     # These tasks should never overlap
@@ -74,9 +76,7 @@ def check_available_tenants(self: Task) -> None:  # noqa: ARG001
             num_available_tenants = db_session.query(AvailableTenant).count()
 
         # Get the target number of available tenants
-        num_minimum_available_tenants = getattr(
-            TARGET_AVAILABLE_TENANTS, "value", DEFAULT_TARGET_AVAILABLE_TENANTS
-        )
+        num_minimum_available_tenants = TARGET_AVAILABLE_TENANTS
 
         # Calculate how many new tenants we need to provision
         if num_available_tenants < num_minimum_available_tenants:
@@ -90,22 +90,87 @@ def check_available_tenants(self: Task) -> None:  # noqa: ARG001
             f"To provision: {tenants_to_provision}"
         )
 
-        # just provision one tenant each time we run this ... increase if needed.
-        if tenants_to_provision > 0:
-            pre_provision_tenant()
+        batch_size = min(tenants_to_provision, _MAX_TENANTS_PER_RUN)
+        if batch_size < tenants_to_provision:
+            task_logger.info(
+                f"Capping batch to {batch_size} (need {tenants_to_provision}, will catch up next cycle)"
+            )
+
+        provisioned = 0
+        for i in range(batch_size):
+            task_logger.info(f"Provisioning tenant {i + 1}/{batch_size}")
+            try:
+                if pre_provision_tenant():
+                    provisioned += 1
+            except Exception:
+                task_logger.exception(
+                    f"Failed to provision tenant {i + 1}/{batch_size}, continuing with remaining tenants"
+                )
+
+        task_logger.info(f"Provisioning complete: {provisioned}/{batch_size} succeeded")
+
+        # Migrate any pool tenants that were provisioned before a new migration was deployed
+        _migrate_stale_pool_tenants()
 
     except Exception:
         task_logger.exception("Error in check_available_tenants task")
 
     finally:
-        lock_check.release()
+        try:
+            lock_check.release()
+        except Exception:
+            task_logger.warning(
+                "Could not release check lock (likely expired), continuing"
+            )
 
 
-def pre_provision_tenant() -> None:
+def _migrate_stale_pool_tenants() -> None:
+    """
+    Run alembic upgrade head on all pool tenants. Since alembic upgrade head is
+    idempotent, tenants already at head are a fast no-op. This ensures pool
+    tenants are always current so that signup doesn't hit schema mismatches
+    (e.g. missing columns added after the tenant was pre-provisioned).
+    """
+    with get_session_with_shared_schema() as db_session:
+        pool_tenants = db_session.query(AvailableTenant).all()
+        tenant_ids = [t.tenant_id for t in pool_tenants]
+
+    if not tenant_ids:
+        return
+
+    task_logger.info(
+        f"Checking {len(tenant_ids)} pool tenant(s) for pending migrations"
+    )
+
+    for tenant_id in tenant_ids:
+        try:
+            run_alembic_migrations(tenant_id)
+            new_version = get_current_alembic_version(tenant_id)
+            with get_session_with_shared_schema() as db_session:
+                tenant = (
+                    db_session.query(AvailableTenant)
+                    .filter_by(tenant_id=tenant_id)
+                    .first()
+                )
+                if tenant and tenant.alembic_version != new_version:
+                    task_logger.info(
+                        f"Migrated pool tenant {tenant_id}: {tenant.alembic_version} -> {new_version}"
+                    )
+                    tenant.alembic_version = new_version
+                    db_session.commit()
+        except Exception:
+            task_logger.exception(
+                f"Failed to migrate pool tenant {tenant_id}, skipping"
+            )
+
+
+def pre_provision_tenant() -> bool:
     """
     Pre-provision a new tenant and store it in the NewAvailableTenant table.
     This function fully sets up the tenant with all necessary configurations,
     so it's ready to be assigned to a user immediately.
+
+    Returns True if a tenant was successfully provisioned, False otherwise.
     """
     # The MULTI_TENANT check is now done at the caller level (check_available_tenants)
     # rather than inside this function
@@ -113,15 +178,15 @@ def pre_provision_tenant() -> None:
     r = get_redis_client(tenant_id=ONYX_CLOUD_TENANT_ID)
     lock_provision: RedisLock = r.lock(
         OnyxRedisLocks.CLOUD_PRE_PROVISION_TENANT_LOCK,
-        timeout=_TENANT_PROVISIONING_SOFT_TIME_LIMIT,
+        timeout=_TENANT_PROVISIONING_TIME_LIMIT,
     )
 
     # Allow multiple pre-provisioning tasks to run, but ensure they don't overlap
     if not lock_provision.acquire(blocking=False):
-        task_logger.debug(
-            "Skipping pre_provision_tenant task because it is already running"
+        task_logger.warning(
+            "Skipping pre_provision_tenant — could not acquire provision lock"
         )
-        return
+        return False
 
     tenant_id: str | None = None
     try:
@@ -161,6 +226,7 @@ def pre_provision_tenant() -> None:
                 db_session.add(new_tenant)
                 db_session.commit()
                 task_logger.info(f"Successfully pre-provisioned tenant: {tenant_id}")
+                return True
             except Exception:
                 db_session.rollback()
                 task_logger.error(
@@ -184,5 +250,11 @@ def pre_provision_tenant() -> None:
                 asyncio.run(rollback_tenant_provisioning(tenant_id))
             except Exception:
                 task_logger.exception(f"Error during rollback for tenant: {tenant_id}")
+        return False
     finally:
-        lock_provision.release()
+        try:
+            lock_provision.release()
+        except Exception:
+            task_logger.warning(
+                "Could not release provision lock (likely expired), continuing"
+            )

backend/ee/onyx/configs/app_configs.py

@@ -118,9 +118,7 @@ JWT_PUBLIC_KEY_URL: str | None = os.getenv("JWT_PUBLIC_KEY_URL", None)
 SUPER_USERS = json.loads(os.environ.get("SUPER_USERS", "[]"))
 SUPER_CLOUD_API_KEY = os.environ.get("SUPER_CLOUD_API_KEY", "api_key")
 
-# The posthog client does not accept empty API keys or hosts however it fails silently
-# when the capture is called. These defaults prevent Posthog issues from breaking the Onyx app
-POSTHOG_API_KEY = os.environ.get("POSTHOG_API_KEY") or "FooBar"
+POSTHOG_API_KEY = os.environ.get("POSTHOG_API_KEY")
 POSTHOG_HOST = os.environ.get("POSTHOG_HOST") or "https://us.i.posthog.com"
 POSTHOG_DEBUG_LOGS_ENABLED = (
     os.environ.get("POSTHOG_DEBUG_LOGS_ENABLED", "").lower() == "true"

backend/ee/onyx/configs/license_enforcement_config.py

@@ -69,5 +69,7 @@ EE_ONLY_PATH_PREFIXES: frozenset[str] = frozenset(
         "/admin/token-rate-limits",
         # Evals
         "/evals",
+        # Hook extensions
+        "/admin/hooks",
     }
 )

backend/ee/onyx/db/token_limit.py

@@ -115,8 +115,14 @@ def fetch_user_group_token_rate_limits_for_user(
     ordered: bool = True,
     get_editable: bool = True,
 ) -> Sequence[TokenRateLimit]:
-    stmt = select(TokenRateLimit)
-    stmt = stmt.where(User__UserGroup.user_group_id == group_id)
+    stmt = (
+        select(TokenRateLimit)
+        .join(
+            TokenRateLimit__UserGroup,
+            TokenRateLimit.id == TokenRateLimit__UserGroup.rate_limit_id,
+        )
+        .where(TokenRateLimit__UserGroup.user_group_id == group_id)
+    )
     stmt = _add_user_filters(stmt, user, get_editable)
 
     if enabled_only:

backend/ee/onyx/db/user_group.py

@@ -800,6 +800,33 @@ def update_user_group(
     return db_user_group
 
 
+def rename_user_group(
+    db_session: Session,
+    user_group_id: int,
+    new_name: str,
+) -> UserGroup:
+    stmt = select(UserGroup).where(UserGroup.id == user_group_id)
+    db_user_group = db_session.scalar(stmt)
+    if db_user_group is None:
+        raise ValueError(f"UserGroup with id '{user_group_id}' not found")
+
+    _check_user_group_is_modifiable(db_user_group)
+
+    db_user_group.name = new_name
+    db_user_group.time_last_modified_by_user = func.now()
+
+    # CC pair documents in Vespa contain the group name, so we need to
+    # trigger a sync to update them with the new name.
+    _mark_user_group__cc_pair_relationships_outdated__no_commit(
+        db_session=db_session, user_group_id=user_group_id
+    )
+    if not DISABLE_VECTOR_DB:
+        db_user_group.is_up_to_date = False
+
+    db_session.commit()
+    return db_user_group
+
+
 def prepare_user_group_for_deletion(db_session: Session, user_group_id: int) -> None:
     stmt = select(UserGroup).where(UserGroup.id == user_group_id)
     db_user_group = db_session.scalar(stmt)

backend/ee/onyx/external_permissions/sharepoint/permission_utils.py

@@ -250,20 +250,24 @@ def _get_sharepoint_list_item_id(drive_item: DriveItem) -> str | None:
         raise e
 
 
-def _is_public_item(drive_item: DriveItem) -> bool:
-    is_public = False
+def _is_public_item(
+    drive_item: DriveItem,
+    treat_sharing_link_as_public: bool = False,
+) -> bool:
+    if not treat_sharing_link_as_public:
+        return False
+
     try:
         permissions = sleep_and_retry(
             drive_item.permissions.get_all(page_loaded=lambda _: None), "is_public_item"
         )
         for permission in permissions:
-            if permission.link and (
-                permission.link.scope == "anonymous"
-                or permission.link.scope == "organization"
+            if permission.link and permission.link.scope in (
+                "anonymous",
+                "organization",
             ):
-                is_public = True
-                break
-        return is_public
+                return True
+        return False
     except Exception as e:
         logger.error(f"Failed to check if item {drive_item.id} is public: {e}")
         return False
@@ -504,6 +508,7 @@ def get_external_access_from_sharepoint(
     drive_item: DriveItem | None,
     site_page: dict[str, Any] | None,
     add_prefix: bool = False,
+    treat_sharing_link_as_public: bool = False,
 ) -> ExternalAccess:
     """
     Get external access information from SharePoint.
@@ -563,8 +568,7 @@ def get_external_access_from_sharepoint(
                     )
 
     if drive_item and drive_name:
-        # Here we check if the item have have any public links, if so we return early
-        is_public = _is_public_item(drive_item)
+        is_public = _is_public_item(drive_item, treat_sharing_link_as_public)
         if is_public:
             logger.info(f"Item {drive_item.id} is public")
             return ExternalAccess(

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

@@ -8,6 +8,7 @@ from ee.onyx.external_permissions.slack.utils import fetch_user_id_to_email_map
 from onyx.access.models import DocExternalAccess
 from onyx.access.models import ExternalAccess
 from onyx.connectors.credentials_provider import OnyxDBCredentialsProvider
+from onyx.connectors.interfaces import SecondsSinceUnixEpoch
 from onyx.connectors.models import HierarchyNode
 from onyx.connectors.slack.connector import get_channels
 from onyx.connectors.slack.connector import make_paginated_slack_api_call
@@ -105,9 +106,11 @@ def _get_slack_document_access(
     slack_connector: SlackConnector,
     channel_permissions: dict[str, ExternalAccess],  # noqa: ARG001
     callback: IndexingHeartbeatInterface | None,
+    indexing_start: SecondsSinceUnixEpoch | None = None,
 ) -> Generator[DocExternalAccess, None, None]:
     slim_doc_generator = slack_connector.retrieve_all_slim_docs_perm_sync(
-        callback=callback
+        callback=callback,
+        start=indexing_start,
     )
 
     for doc_metadata_batch in slim_doc_generator:
@@ -180,9 +183,15 @@ def slack_doc_sync(
 
     slack_connector = SlackConnector(**cc_pair.connector.connector_specific_config)
     slack_connector.set_credentials_provider(provider)
+    indexing_start_ts: SecondsSinceUnixEpoch | None = (
+        cc_pair.connector.indexing_start.timestamp()
+        if cc_pair.connector.indexing_start is not None
+        else None
+    )
 
     yield from _get_slack_document_access(
-        slack_connector,
+        slack_connector=slack_connector,
         channel_permissions=channel_permissions,
         callback=callback,
+        indexing_start=indexing_start_ts,
     )

backend/ee/onyx/external_permissions/utils.py

@@ -6,6 +6,7 @@ from onyx.access.models import ElementExternalAccess
 from onyx.access.models import ExternalAccess
 from onyx.access.models import NodeExternalAccess
 from onyx.configs.constants import DocumentSource
+from onyx.connectors.interfaces import SecondsSinceUnixEpoch
 from onyx.connectors.interfaces import SlimConnectorWithPermSync
 from onyx.connectors.models import HierarchyNode
 from onyx.db.models import ConnectorCredentialPair
@@ -40,10 +41,19 @@ def generic_doc_sync(
 
     logger.info(f"Starting {doc_source} doc sync for CC Pair ID: {cc_pair.id}")
 
+    indexing_start: SecondsSinceUnixEpoch | None = (
+        cc_pair.connector.indexing_start.timestamp()
+        if cc_pair.connector.indexing_start is not None
+        else None
+    )
+
     newly_fetched_doc_ids: set[str] = set()
 
     logger.info(f"Fetching all slim documents from {doc_source}")
-    for doc_batch in slim_connector.retrieve_all_slim_docs_perm_sync(callback=callback):
+    for doc_batch in slim_connector.retrieve_all_slim_docs_perm_sync(
+        start=indexing_start,
+        callback=callback,
+    ):
         logger.info(f"Got {len(doc_batch)} slim documents from {doc_source}")
 
         if callback:

backend/ee/onyx/feature_flags/posthog_provider.py

@@ -34,6 +34,9 @@ class PostHogFeatureFlagProvider(FeatureFlagProvider):
         Returns:
             True if the feature is enabled for the user, False otherwise.
         """
+        if not posthog:
+            return False
+
         try:
             posthog.set(
                 distinct_id=user_id,

backend/ee/onyx/hooks/executor.py

@@ -0,0 +1,385 @@
+"""Hook executor — calls a customer's external HTTP endpoint for a given hook point.
+
+Usage (Celery tasks and FastAPI handlers):
+    result = execute_hook(
+        db_session=db_session,
+        hook_point=HookPoint.QUERY_PROCESSING,
+        payload={"query": "...", "user_email": "...", "chat_session_id": "..."},
+        response_type=QueryProcessingResponse,
+    )
+
+    if isinstance(result, HookSkipped):
+        # no active hook configured — continue with original behavior
+        ...
+    elif isinstance(result, HookSoftFailed):
+        # hook failed but fail strategy is SOFT — continue with original behavior
+        ...
+    else:
+        # result is a validated Pydantic model instance (response_type)
+        ...
+
+is_reachable update policy
+--------------------------
+``is_reachable`` on the Hook row is updated selectively — only when the outcome
+carries meaningful signal about physical reachability:
+
+  NetworkError (DNS, connection refused)  → False  (cannot reach the server)
+  HTTP 401 / 403                          → False  (api_key revoked or invalid)
+  TimeoutException                        → None   (server may be slow, skip write)
+  Other HTTP errors (4xx / 5xx)           → None   (server responded, skip write)
+  Unknown exception                       → None   (no signal, skip write)
+  Non-JSON / non-dict response            → None   (server responded, skip write)
+  Success (2xx, valid dict)               → True   (confirmed reachable)
+
+None means "leave the current value unchanged" — no DB round-trip is made.
+
+DB session design
+-----------------
+The executor uses three sessions:
+
+  1. Caller's session (db_session) — used only for the hook lookup read. All
+     needed fields are extracted from the Hook object before the HTTP call, so
+     the caller's session is not held open during the external HTTP request.
+
+  2. Log session — a separate short-lived session opened after the HTTP call
+     completes to write the HookExecutionLog row on failure. Success runs are
+     not recorded. Committed independently of everything else.
+
+  3. Reachable session — a second short-lived session to update is_reachable on
+     the Hook. Kept separate from the log session so a concurrent hook deletion
+     (which causes update_hook__no_commit to raise OnyxError(NOT_FOUND)) cannot
+     prevent the execution log from being written. This update is best-effort.
+"""
+
+import json
+import time
+from typing import Any
+from typing import TypeVar
+
+import httpx
+from pydantic import BaseModel
+from pydantic import ValidationError
+from sqlalchemy.orm import Session
+
+from onyx.db.engine.sql_engine import get_session_with_current_tenant
+from onyx.db.enums import HookFailStrategy
+from onyx.db.enums import HookPoint
+from onyx.db.hook import create_hook_execution_log__no_commit
+from onyx.db.hook import get_non_deleted_hook_by_hook_point
+from onyx.db.hook import update_hook__no_commit
+from onyx.db.models import Hook
+from onyx.error_handling.error_codes import OnyxErrorCode
+from onyx.error_handling.exceptions import OnyxError
+from onyx.hooks.executor import HookSkipped
+from onyx.hooks.executor import HookSoftFailed
+from onyx.utils.logger import setup_logger
+from shared_configs.configs import MULTI_TENANT
+
+logger = setup_logger()
+
+
+T = TypeVar("T", bound=BaseModel)
+
+
+# ---------------------------------------------------------------------------
+# Private helpers
+# ---------------------------------------------------------------------------
+
+
+class _HttpOutcome(BaseModel):
+    """Structured result of an HTTP hook call, returned by _process_response."""
+
+    is_success: bool
+    updated_is_reachable: (
+        bool | None
+    )  # True/False = write to DB, None = unchanged (skip write)
+    status_code: int | None
+    error_message: str | None
+    response_payload: dict[str, Any] | None
+
+
+def _lookup_hook(
+    db_session: Session,
+    hook_point: HookPoint,
+) -> Hook | HookSkipped:
+    """Return the active Hook or HookSkipped if hooks are unavailable/unconfigured.
+
+    No HTTP call is made and no DB writes are performed for any HookSkipped path.
+    There is nothing to log and no reachability information to update.
+    """
+    if MULTI_TENANT:
+        return HookSkipped()
+    hook = get_non_deleted_hook_by_hook_point(
+        db_session=db_session, hook_point=hook_point
+    )
+    if hook is None or not hook.is_active:
+        return HookSkipped()
+    if not hook.endpoint_url:
+        return HookSkipped()
+    return hook
+
+
+def _process_response(
+    *,
+    response: httpx.Response | None,
+    exc: Exception | None,
+    timeout: float,
+) -> _HttpOutcome:
+    """Process the result of an HTTP call and return a structured outcome.
+
+    Called after the client.post() try/except. If post() raised, exc is set and
+    response is None. Otherwise response is set and exc is None. Handles
+    raise_for_status(), JSON decoding, and the dict shape check.
+    """
+    if exc is not None:
+        if isinstance(exc, httpx.NetworkError):
+            msg = f"Hook network error (endpoint unreachable): {exc}"
+            logger.warning(msg, exc_info=exc)
+            return _HttpOutcome(
+                is_success=False,
+                updated_is_reachable=False,
+                status_code=None,
+                error_message=msg,
+                response_payload=None,
+            )
+        if isinstance(exc, httpx.TimeoutException):
+            msg = f"Hook timed out after {timeout}s: {exc}"
+            logger.warning(msg, exc_info=exc)
+            return _HttpOutcome(
+                is_success=False,
+                updated_is_reachable=None,  # timeout doesn't indicate unreachability
+                status_code=None,
+                error_message=msg,
+                response_payload=None,
+            )
+        msg = f"Hook call failed: {exc}"
+        logger.exception(msg, exc_info=exc)
+        return _HttpOutcome(
+            is_success=False,
+            updated_is_reachable=None,  # unknown error — don't make assumptions
+            status_code=None,
+            error_message=msg,
+            response_payload=None,
+        )
+
+    if response is None:
+        raise ValueError(
+            "exactly one of response or exc must be non-None; both are None"
+        )
+    status_code = response.status_code
+
+    try:
+        response.raise_for_status()
+    except httpx.HTTPStatusError as e:
+        msg = f"Hook returned HTTP {e.response.status_code}: {e.response.text}"
+        logger.warning(msg, exc_info=e)
+        # 401/403 means the api_key has been revoked or is invalid — mark unreachable
+        # so the operator knows to update it. All other HTTP errors keep is_reachable
+        # as-is (server is up, the request just failed for application reasons).
+        auth_failed = e.response.status_code in (401, 403)
+        return _HttpOutcome(
+            is_success=False,
+            updated_is_reachable=False if auth_failed else None,
+            status_code=status_code,
+            error_message=msg,
+            response_payload=None,
+        )
+
+    try:
+        response_payload = response.json()
+    except (json.JSONDecodeError, httpx.DecodingError) as e:
+        msg = f"Hook returned non-JSON response: {e}"
+        logger.warning(msg, exc_info=e)
+        return _HttpOutcome(
+            is_success=False,
+            updated_is_reachable=None,  # server responded — reachability unchanged
+            status_code=status_code,
+            error_message=msg,
+            response_payload=None,
+        )
+
+    if not isinstance(response_payload, dict):
+        msg = f"Hook returned non-dict JSON (got {type(response_payload).__name__})"
+        logger.warning(msg)
+        return _HttpOutcome(
+            is_success=False,
+            updated_is_reachable=None,  # server responded — reachability unchanged
+            status_code=status_code,
+            error_message=msg,
+            response_payload=None,
+        )
+
+    return _HttpOutcome(
+        is_success=True,
+        updated_is_reachable=True,
+        status_code=status_code,
+        error_message=None,
+        response_payload=response_payload,
+    )
+
+
+def _persist_result(
+    *,
+    hook_id: int,
+    outcome: _HttpOutcome,
+    duration_ms: int,
+) -> None:
+    """Write the execution log on failure and optionally update is_reachable, each
+    in its own session so a failure in one does not affect the other."""
+    # Only write the execution log on failure — success runs are not recorded.
+    # Must not be skipped if the is_reachable update fails (e.g. hook concurrently
+    # deleted between the initial lookup and here).
+    if not outcome.is_success:
+        try:
+            with get_session_with_current_tenant() as log_session:
+                create_hook_execution_log__no_commit(
+                    db_session=log_session,
+                    hook_id=hook_id,
+                    is_success=False,
+                    error_message=outcome.error_message,
+                    status_code=outcome.status_code,
+                    duration_ms=duration_ms,
+                )
+                log_session.commit()
+        except Exception:
+            logger.exception(
+                f"Failed to persist hook execution log for hook_id={hook_id}"
+            )
+
+    # Update is_reachable separately — best-effort, non-critical.
+    # None means the value is unchanged (set by the caller to skip the no-op write).
+    # update_hook__no_commit can raise OnyxError(NOT_FOUND) if the hook was
+    # concurrently deleted, so keep this isolated from the log write above.
+    if outcome.updated_is_reachable is not None:
+        try:
+            with get_session_with_current_tenant() as reachable_session:
+                update_hook__no_commit(
+                    db_session=reachable_session,
+                    hook_id=hook_id,
+                    is_reachable=outcome.updated_is_reachable,
+                )
+                reachable_session.commit()
+        except Exception:
+            logger.warning(f"Failed to update is_reachable for hook_id={hook_id}")
+
+
+# ---------------------------------------------------------------------------
+# Public API
+# ---------------------------------------------------------------------------
+
+
+def _execute_hook_inner(
+    hook: Hook,
+    payload: dict[str, Any],
+    response_type: type[T],
+) -> T | HookSoftFailed:
+    """Make the HTTP call, validate the response, and return a typed model.
+
+    Raises OnyxError on HARD failure. Returns HookSoftFailed on SOFT failure.
+    """
+    timeout = hook.timeout_seconds
+    hook_id = hook.id
+    fail_strategy = hook.fail_strategy
+    endpoint_url = hook.endpoint_url
+    current_is_reachable: bool | None = hook.is_reachable
+
+    if not endpoint_url:
+        raise ValueError(
+            f"hook_id={hook_id} is active but has no endpoint_url — "
+            "active hooks without an endpoint_url must be rejected by _lookup_hook"
+        )
+
+    start = time.monotonic()
+    response: httpx.Response | None = None
+    exc: Exception | None = None
+    try:
+        api_key: str | None = (
+            hook.api_key.get_value(apply_mask=False) if hook.api_key else None
+        )
+        headers: dict[str, str] = {"Content-Type": "application/json"}
+        if api_key:
+            headers["Authorization"] = f"Bearer {api_key}"
+        with httpx.Client(
+            timeout=timeout, follow_redirects=False
+        ) as client:  # SSRF guard: never follow redirects
+            response = client.post(endpoint_url, json=payload, headers=headers)
+    except Exception as e:
+        exc = e
+    duration_ms = int((time.monotonic() - start) * 1000)
+
+    outcome = _process_response(response=response, exc=exc, timeout=timeout)
+
+    # Validate the response payload against response_type.
+    # A validation failure downgrades the outcome to a failure so it is logged,
+    # is_reachable is left unchanged (server responded — just a bad payload),
+    # and fail_strategy is respected below.
+    validated_model: T | None = None
+    if outcome.is_success and outcome.response_payload is not None:
+        try:
+            validated_model = response_type.model_validate(outcome.response_payload)
+        except ValidationError as e:
+            msg = (
+                f"Hook response failed validation against {response_type.__name__}: {e}"
+            )
+            outcome = _HttpOutcome(
+                is_success=False,
+                updated_is_reachable=None,  # server responded — reachability unchanged
+                status_code=outcome.status_code,
+                error_message=msg,
+                response_payload=None,
+            )
+
+    # Skip the is_reachable write when the value would not change — avoids a
+    # no-op DB round-trip on every call when the hook is already in the expected state.
+    if outcome.updated_is_reachable == current_is_reachable:
+        outcome = outcome.model_copy(update={"updated_is_reachable": None})
+    _persist_result(hook_id=hook_id, outcome=outcome, duration_ms=duration_ms)
+
+    if not outcome.is_success:
+        if fail_strategy == HookFailStrategy.HARD:
+            raise OnyxError(
+                OnyxErrorCode.HOOK_EXECUTION_FAILED,
+                outcome.error_message or "Hook execution failed.",
+            )
+        logger.warning(
+            f"Hook execution failed (soft fail) for hook_id={hook_id}: {outcome.error_message}"
+        )
+        return HookSoftFailed()
+
+    if validated_model is None:
+        raise OnyxError(
+            OnyxErrorCode.INTERNAL_ERROR,
+            f"validated_model is None for successful hook call (hook_id={hook_id})",
+        )
+    return validated_model
+
+
+def _execute_hook_impl(
+    *,
+    db_session: Session,
+    hook_point: HookPoint,
+    payload: dict[str, Any],
+    response_type: type[T],
+) -> T | HookSkipped | HookSoftFailed:
+    """EE implementation — loaded by CE's execute_hook via fetch_versioned_implementation.
+
+    Returns HookSkipped if no active hook is configured, HookSoftFailed if the
+    hook failed with SOFT fail strategy, or a validated response model on success.
+    Raises OnyxError on HARD failure or if the hook is misconfigured.
+    """
+    hook = _lookup_hook(db_session, hook_point)
+    if isinstance(hook, HookSkipped):
+        return hook
+
+    fail_strategy = hook.fail_strategy
+    hook_id = hook.id
+
+    try:
+        return _execute_hook_inner(hook, payload, response_type)
+    except Exception:
+        if fail_strategy == HookFailStrategy.SOFT:
+            logger.exception(
+                f"Unexpected error in hook execution (soft fail) for hook_id={hook_id}"
+            )
+            return HookSoftFailed()
+        raise

backend/ee/onyx/main.py

@@ -15,6 +15,7 @@ from ee.onyx.server.enterprise_settings.api import (
     basic_router as enterprise_settings_router,
 )
 from ee.onyx.server.evals.api import router as evals_router
+from ee.onyx.server.features.hooks.api import router as hook_router
 from ee.onyx.server.license.api import router as license_router
 from ee.onyx.server.manage.standard_answer import router as standard_answer_router
 from ee.onyx.server.middleware.license_enforcement import (
@@ -138,6 +139,7 @@ def get_application() -> FastAPI:
     include_router_with_global_prefix_prepended(application, ee_oauth_router)
     include_router_with_global_prefix_prepended(application, ee_document_cc_pair_router)
     include_router_with_global_prefix_prepended(application, evals_router)
+    include_router_with_global_prefix_prepended(application, hook_router)
 
     # Enterprise-only global settings
     include_router_with_global_prefix_prepended(

backend/ee/onyx/search/process_search_query.py

@@ -44,19 +44,21 @@ def _run_single_search(
     user: User,
     db_session: Session,
     num_hits: int | None = None,
+    hybrid_alpha: float | None = None,
 ) -> list[InferenceChunk]:
     """Execute a single search query and return chunks."""
     chunk_search_request = ChunkSearchRequest(
         query=query,
         user_selected_filters=filters,
         limit=num_hits,
+        hybrid_alpha=hybrid_alpha,
     )
 
     return search_pipeline(
         chunk_search_request=chunk_search_request,
         document_index=document_index,
         user=user,
-        persona=None,  # No persona for direct search
+        persona_search_info=None,
         db_session=db_session,
     )
 
@@ -74,7 +76,7 @@ def stream_search_query(
     Core search function that yields streaming packets.
     Used by both streaming and non-streaming endpoints.
     """
-    # Get document index
+    # Get document index.
     search_settings = get_current_search_settings(db_session)
     # This flow is for search so we do not get all indices.
     document_index = get_default_document_index(search_settings, None, db_session)
@@ -119,6 +121,7 @@ def stream_search_query(
             user=user,
             db_session=db_session,
             num_hits=request.num_hits,
+            hybrid_alpha=request.hybrid_alpha,
         )
     else:
         # Multiple queries - run in parallel and merge with RRF
@@ -133,6 +136,7 @@ def stream_search_query(
                     user,
                     db_session,
                     request.num_hits,
+                    request.hybrid_alpha,
                 ),
             )
             for query in all_executed_queries

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

@@ -157,7 +157,11 @@ def fetch_logo_helper(db_session: Session) -> Response:  # noqa: ARG001
             detail="No logo file found",
         )
     else:
-        return Response(content=onyx_file.data, media_type=onyx_file.mime_type)
+        return Response(
+            content=onyx_file.data,
+            media_type=onyx_file.mime_type,
+            headers={"Cache-Control": "no-cache"},
+        )
 
 
 def fetch_logotype_helper(db_session: Session) -> Response:  # noqa: ARG001

backend/ee/onyx/server/features/hooks/api.py

@@ -0,0 +1,449 @@
+import httpx
+from fastapi import APIRouter
+from fastapi import Depends
+from fastapi import Query
+from sqlalchemy.orm import Session
+
+from onyx.auth.users import current_admin_user
+from onyx.auth.users import User
+from onyx.db.constants import UNSET
+from onyx.db.constants import UnsetType
+from onyx.db.engine.sql_engine import get_session
+from onyx.db.engine.sql_engine import get_session_with_current_tenant
+from onyx.db.hook import create_hook__no_commit
+from onyx.db.hook import delete_hook__no_commit
+from onyx.db.hook import get_hook_by_id
+from onyx.db.hook import get_hook_execution_logs
+from onyx.db.hook import get_hooks
+from onyx.db.hook import update_hook__no_commit
+from onyx.db.models import Hook
+from onyx.error_handling.error_codes import OnyxErrorCode
+from onyx.error_handling.exceptions import OnyxError
+from onyx.hooks.api_dependencies import require_hook_enabled
+from onyx.hooks.models import HookCreateRequest
+from onyx.hooks.models import HookExecutionRecord
+from onyx.hooks.models import HookPointMetaResponse
+from onyx.hooks.models import HookResponse
+from onyx.hooks.models import HookUpdateRequest
+from onyx.hooks.models import HookValidateResponse
+from onyx.hooks.models import HookValidateStatus
+from onyx.hooks.registry import get_all_specs
+from onyx.hooks.registry import get_hook_point_spec
+from onyx.utils.logger import setup_logger
+from onyx.utils.url import SSRFException
+from onyx.utils.url import validate_outbound_http_url
+
+logger = setup_logger()
+
+# ---------------------------------------------------------------------------
+# SSRF protection
+# ---------------------------------------------------------------------------
+
+
+def _check_ssrf_safety(endpoint_url: str) -> None:
+    """Raise OnyxError if endpoint_url could be used for SSRF.
+
+    Delegates to validate_outbound_http_url with https_only=True.
+    Uses BAD_GATEWAY so the frontend maps the error to the Endpoint URL field.
+    """
+    try:
+        validate_outbound_http_url(endpoint_url, https_only=True)
+    except (SSRFException, ValueError) as e:
+        raise OnyxError(OnyxErrorCode.BAD_GATEWAY, str(e))
+
+
+# ---------------------------------------------------------------------------
+# Helpers
+# ---------------------------------------------------------------------------
+
+
+def _hook_to_response(hook: Hook, creator_email: str | None = None) -> HookResponse:
+    return HookResponse(
+        id=hook.id,
+        name=hook.name,
+        hook_point=hook.hook_point,
+        endpoint_url=hook.endpoint_url,
+        api_key_masked=(
+            hook.api_key.get_value(apply_mask=True) if hook.api_key else None
+        ),
+        fail_strategy=hook.fail_strategy,
+        timeout_seconds=hook.timeout_seconds,
+        is_active=hook.is_active,
+        is_reachable=hook.is_reachable,
+        creator_email=(
+            creator_email
+            if creator_email is not None
+            else (hook.creator.email if hook.creator else None)
+        ),
+        created_at=hook.created_at,
+        updated_at=hook.updated_at,
+    )
+
+
+def _get_hook_or_404(
+    db_session: Session,
+    hook_id: int,
+    include_creator: bool = False,
+) -> Hook:
+    hook = get_hook_by_id(
+        db_session=db_session,
+        hook_id=hook_id,
+        include_creator=include_creator,
+    )
+    if hook is None:
+        raise OnyxError(OnyxErrorCode.NOT_FOUND, f"Hook {hook_id} not found.")
+    return hook
+
+
+def _raise_for_validation_failure(validation: HookValidateResponse) -> None:
+    """Raise an appropriate OnyxError for a non-passed validation result."""
+    if validation.status == HookValidateStatus.auth_failed:
+        raise OnyxError(OnyxErrorCode.CREDENTIAL_INVALID, validation.error_message)
+    if validation.status == HookValidateStatus.timeout:
+        raise OnyxError(
+            OnyxErrorCode.GATEWAY_TIMEOUT,
+            f"Endpoint validation failed: {validation.error_message}",
+        )
+    raise OnyxError(
+        OnyxErrorCode.BAD_GATEWAY,
+        f"Endpoint validation failed: {validation.error_message}",
+    )
+
+
+def _validate_endpoint(
+    endpoint_url: str,
+    api_key: str | None,
+    timeout_seconds: float,
+) -> HookValidateResponse:
+    """Check whether endpoint_url is reachable by sending an empty POST request.
+
+    We use POST since hook endpoints expect POST requests. The server will typically
+    respond with 4xx (missing/invalid body) — that is fine. Any HTTP response means
+    the server is up and routable. A 401/403 response returns auth_failed
+    (not reachable — indicates the api_key is invalid).
+
+    Timeout handling:
+    - Any httpx.TimeoutException (ConnectTimeout, ReadTimeout, WriteTimeout, PoolTimeout) →
+      timeout (operator should consider increasing timeout_seconds).
+    - All other exceptions → cannot_connect.
+    """
+    _check_ssrf_safety(endpoint_url)
+    headers: dict[str, str] = {}
+    if api_key:
+        headers["Authorization"] = f"Bearer {api_key}"
+    try:
+        with httpx.Client(timeout=timeout_seconds, follow_redirects=False) as client:
+            response = client.post(endpoint_url, headers=headers)
+        if response.status_code in (401, 403):
+            return HookValidateResponse(
+                status=HookValidateStatus.auth_failed,
+                error_message=f"Authentication failed (HTTP {response.status_code})",
+            )
+        return HookValidateResponse(status=HookValidateStatus.passed)
+    except httpx.TimeoutException as exc:
+        # Any timeout (connect, read, or write) means the configured timeout_seconds
+        # is too low for this endpoint. Report as timeout so the UI directs the user
+        # to increase the timeout setting.
+        logger.warning(
+            "Hook endpoint validation: timeout for %s",
+            endpoint_url,
+            exc_info=exc,
+        )
+        return HookValidateResponse(
+            status=HookValidateStatus.timeout,
+            error_message="Endpoint timed out — consider increasing timeout_seconds.",
+        )
+    except Exception as exc:
+        logger.warning(
+            "Hook endpoint validation: connection error for %s",
+            endpoint_url,
+            exc_info=exc,
+        )
+        return HookValidateResponse(
+            status=HookValidateStatus.cannot_connect, error_message=str(exc)
+        )
+
+
+# ---------------------------------------------------------------------------
+# Routers
+# ---------------------------------------------------------------------------
+
+router = APIRouter(prefix="/admin/hooks")
+
+
+# ---------------------------------------------------------------------------
+# Hook endpoints
+# ---------------------------------------------------------------------------
+
+
+@router.get("/specs")
+def get_hook_point_specs(
+    _: User = Depends(current_admin_user),
+    _hook_enabled: None = Depends(require_hook_enabled),
+) -> list[HookPointMetaResponse]:
+    return [
+        HookPointMetaResponse(
+            hook_point=spec.hook_point,
+            display_name=spec.display_name,
+            description=spec.description,
+            docs_url=spec.docs_url,
+            input_schema=spec.input_schema,
+            output_schema=spec.output_schema,
+            default_timeout_seconds=spec.default_timeout_seconds,
+            default_fail_strategy=spec.default_fail_strategy,
+            fail_hard_description=spec.fail_hard_description,
+        )
+        for spec in get_all_specs()
+    ]
+
+
+@router.get("")
+def list_hooks(
+    _: User = Depends(current_admin_user),
+    _hook_enabled: None = Depends(require_hook_enabled),
+    db_session: Session = Depends(get_session),
+) -> list[HookResponse]:
+    hooks = get_hooks(db_session=db_session, include_creator=True)
+    return [_hook_to_response(h) for h in hooks]
+
+
+@router.post("")
+def create_hook(
+    req: HookCreateRequest,
+    user: User = Depends(current_admin_user),
+    _hook_enabled: None = Depends(require_hook_enabled),
+    db_session: Session = Depends(get_session),
+) -> HookResponse:
+    """Create a new hook. The endpoint is validated before persisting — creation fails if
+    the endpoint cannot be reached or the api_key is invalid. Hooks are created active.
+    """
+    spec = get_hook_point_spec(req.hook_point)
+    api_key = req.api_key.get_secret_value() if req.api_key else None
+    validation = _validate_endpoint(
+        endpoint_url=req.endpoint_url,
+        api_key=api_key,
+        timeout_seconds=req.timeout_seconds or spec.default_timeout_seconds,
+    )
+    if validation.status != HookValidateStatus.passed:
+        _raise_for_validation_failure(validation)
+
+    hook = create_hook__no_commit(
+        db_session=db_session,
+        name=req.name,
+        hook_point=req.hook_point,
+        endpoint_url=req.endpoint_url,
+        api_key=api_key,
+        fail_strategy=req.fail_strategy or spec.default_fail_strategy,
+        timeout_seconds=req.timeout_seconds or spec.default_timeout_seconds,
+        is_active=True,
+        is_reachable=True,
+        creator_id=user.id,
+    )
+    db_session.commit()
+    return _hook_to_response(hook, creator_email=user.email)
+
+
+@router.get("/{hook_id}")
+def get_hook(
+    hook_id: int,
+    _: User = Depends(current_admin_user),
+    _hook_enabled: None = Depends(require_hook_enabled),
+    db_session: Session = Depends(get_session),
+) -> HookResponse:
+    hook = _get_hook_or_404(db_session, hook_id, include_creator=True)
+    return _hook_to_response(hook)
+
+
+@router.patch("/{hook_id}")
+def update_hook(
+    hook_id: int,
+    req: HookUpdateRequest,
+    _: User = Depends(current_admin_user),
+    _hook_enabled: None = Depends(require_hook_enabled),
+    db_session: Session = Depends(get_session),
+) -> HookResponse:
+    """Update hook fields. If endpoint_url, api_key, or timeout_seconds changes, the
+    endpoint is re-validated using the effective values. For active hooks the update is
+    rejected on validation failure, keeping live traffic unaffected. For inactive hooks
+    the update goes through regardless and is_reachable is updated to reflect the result.
+
+    Note: if an active hook's endpoint is currently down, even a timeout_seconds-only
+    increase will be rejected. The recovery flow is: deactivate → update → reactivate.
+    """
+    # api_key: UNSET = no change, None = clear, value = update
+    api_key: str | None | UnsetType
+    if "api_key" not in req.model_fields_set:
+        api_key = UNSET
+    elif req.api_key is None:
+        api_key = None
+    else:
+        api_key = req.api_key.get_secret_value()
+
+    endpoint_url_changing = "endpoint_url" in req.model_fields_set
+    api_key_changing = not isinstance(api_key, UnsetType)
+    timeout_changing = "timeout_seconds" in req.model_fields_set
+
+    validated_is_reachable: bool | None = None
+    if endpoint_url_changing or api_key_changing or timeout_changing:
+        existing = _get_hook_or_404(db_session, hook_id)
+        effective_url: str = (
+            req.endpoint_url if endpoint_url_changing else existing.endpoint_url  # type: ignore[assignment]  # endpoint_url is required on create and cannot be cleared on update
+        )
+        effective_api_key: str | None = (
+            (api_key if not isinstance(api_key, UnsetType) else None)
+            if api_key_changing
+            else (
+                existing.api_key.get_value(apply_mask=False)
+                if existing.api_key
+                else None
+            )
+        )
+        effective_timeout: float = (
+            req.timeout_seconds if timeout_changing else existing.timeout_seconds  # type: ignore[assignment]  # req.timeout_seconds is non-None when timeout_changing (validated by HookUpdateRequest)
+        )
+        validation = _validate_endpoint(
+            endpoint_url=effective_url,
+            api_key=effective_api_key,
+            timeout_seconds=effective_timeout,
+        )
+        if existing.is_active and validation.status != HookValidateStatus.passed:
+            _raise_for_validation_failure(validation)
+        validated_is_reachable = validation.status == HookValidateStatus.passed
+
+    hook = update_hook__no_commit(
+        db_session=db_session,
+        hook_id=hook_id,
+        name=req.name,
+        endpoint_url=(req.endpoint_url if endpoint_url_changing else UNSET),
+        api_key=api_key,
+        fail_strategy=req.fail_strategy,
+        timeout_seconds=req.timeout_seconds,
+        is_reachable=validated_is_reachable,
+        include_creator=True,
+    )
+    db_session.commit()
+    return _hook_to_response(hook)
+
+
+@router.delete("/{hook_id}")
+def delete_hook(
+    hook_id: int,
+    _: User = Depends(current_admin_user),
+    _hook_enabled: None = Depends(require_hook_enabled),
+    db_session: Session = Depends(get_session),
+) -> None:
+    delete_hook__no_commit(db_session=db_session, hook_id=hook_id)
+    db_session.commit()
+
+
+@router.post("/{hook_id}/activate")
+def activate_hook(
+    hook_id: int,
+    _: User = Depends(current_admin_user),
+    _hook_enabled: None = Depends(require_hook_enabled),
+    db_session: Session = Depends(get_session),
+) -> HookResponse:
+    hook = _get_hook_or_404(db_session, hook_id)
+    if not hook.endpoint_url:
+        raise OnyxError(
+            OnyxErrorCode.INVALID_INPUT, "Hook has no endpoint URL configured."
+        )
+
+    api_key = hook.api_key.get_value(apply_mask=False) if hook.api_key else None
+    validation = _validate_endpoint(
+        endpoint_url=hook.endpoint_url,
+        api_key=api_key,
+        timeout_seconds=hook.timeout_seconds,
+    )
+    if validation.status != HookValidateStatus.passed:
+        # Persist is_reachable=False in a separate session so the request
+        # session has no commits on the failure path and the transaction
+        # boundary stays clean.
+        if hook.is_reachable is not False:
+            with get_session_with_current_tenant() as side_session:
+                update_hook__no_commit(
+                    db_session=side_session, hook_id=hook_id, is_reachable=False
+                )
+                side_session.commit()
+        _raise_for_validation_failure(validation)
+
+    hook = update_hook__no_commit(
+        db_session=db_session,
+        hook_id=hook_id,
+        is_active=True,
+        is_reachable=True,
+        include_creator=True,
+    )
+    db_session.commit()
+    return _hook_to_response(hook)
+
+
+@router.post("/{hook_id}/validate")
+def validate_hook(
+    hook_id: int,
+    _: User = Depends(current_admin_user),
+    _hook_enabled: None = Depends(require_hook_enabled),
+    db_session: Session = Depends(get_session),
+) -> HookValidateResponse:
+    hook = _get_hook_or_404(db_session, hook_id)
+    if not hook.endpoint_url:
+        raise OnyxError(
+            OnyxErrorCode.INVALID_INPUT, "Hook has no endpoint URL configured."
+        )
+
+    api_key = hook.api_key.get_value(apply_mask=False) if hook.api_key else None
+    validation = _validate_endpoint(
+        endpoint_url=hook.endpoint_url,
+        api_key=api_key,
+        timeout_seconds=hook.timeout_seconds,
+    )
+    validation_passed = validation.status == HookValidateStatus.passed
+    if hook.is_reachable != validation_passed:
+        update_hook__no_commit(
+            db_session=db_session, hook_id=hook_id, is_reachable=validation_passed
+        )
+        db_session.commit()
+    return validation
+
+
+@router.post("/{hook_id}/deactivate")
+def deactivate_hook(
+    hook_id: int,
+    _: User = Depends(current_admin_user),
+    _hook_enabled: None = Depends(require_hook_enabled),
+    db_session: Session = Depends(get_session),
+) -> HookResponse:
+    hook = update_hook__no_commit(
+        db_session=db_session,
+        hook_id=hook_id,
+        is_active=False,
+        include_creator=True,
+    )
+    db_session.commit()
+    return _hook_to_response(hook)
+
+
+# ---------------------------------------------------------------------------
+# Execution log endpoints
+# ---------------------------------------------------------------------------
+
+
+@router.get("/{hook_id}/execution-logs")
+def list_hook_execution_logs(
+    hook_id: int,
+    limit: int = Query(default=10, ge=1, le=100),
+    _: User = Depends(current_admin_user),
+    _hook_enabled: None = Depends(require_hook_enabled),
+    db_session: Session = Depends(get_session),
+) -> list[HookExecutionRecord]:
+    _get_hook_or_404(db_session, hook_id)
+    logs = get_hook_execution_logs(db_session=db_session, hook_id=hook_id, limit=limit)
+    return [
+        HookExecutionRecord(
+            error_message=log.error_message,
+            status_code=log.status_code,
+            duration_ms=log.duration_ms,
+            created_at=log.created_at,
+        )
+        for log in logs
+    ]

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

@@ -27,15 +27,17 @@ class SearchFlowClassificationResponse(BaseModel):
     is_search_flow: bool
 
 
-# NOTE: This model is used for the core flow of the Onyx application, any changes to it should be reviewed and approved by an
-# experienced team member. It is very important to 1. avoid bloat and 2. that this remains backwards compatible across versions.
+# NOTE: This model is used for the core flow of the Onyx application, any
+# changes to it should be reviewed and approved by an experienced team member.
+# It is very important to 1. avoid bloat and 2. that this remains backwards
+# compatible across versions.
 class SendSearchQueryRequest(BaseModel):
     search_query: str
     filters: BaseFilters | None = None
     num_docs_fed_to_llm_selection: int | None = None
     run_query_expansion: bool = False
     num_hits: int = 30
-
+    hybrid_alpha: float | None = None
     include_content: bool = False
     stream: bool = False
 

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

@@ -20,6 +20,7 @@ from ee.onyx.server.query_and_chat.models import SearchQueryResponse
 from ee.onyx.server.query_and_chat.models import SendSearchQueryRequest
 from ee.onyx.server.query_and_chat.streaming_models import SearchErrorPacket
 from onyx.auth.users import current_user
+from onyx.configs.app_configs import ONYX_SEARCH_UI_USES_OPENSEARCH_KEYWORD_SEARCH
 from onyx.db.engine.sql_engine import get_session
 from onyx.db.engine.sql_engine import get_session_with_current_tenant
 from onyx.db.models import User
@@ -67,8 +68,10 @@ def search_flow_classification(
     return SearchFlowClassificationResponse(is_search_flow=is_search_flow)
 
 
-# NOTE: This endpoint is used for the core flow of the Onyx application, any changes to it should be reviewed and approved by an
-# experienced team member. It is very important to 1. avoid bloat and 2. that this remains backwards compatible across versions.
+# NOTE: This endpoint is used for the core flow of the Onyx application, any
+# changes to it should be reviewed and approved by an experienced team member.
+# It is very important to 1. avoid bloat and 2. that this remains backwards
+# compatible across versions.
 @router.post(
     "/send-search-message",
     response_model=None,
@@ -80,13 +83,19 @@ def handle_send_search_message(
     db_session: Session = Depends(get_session),
 ) -> StreamingResponse | SearchFullResponse:
     """
-    Execute a search query with optional streaming.
+    Executes a search query with optional streaming.
 
-    When stream=True: Returns StreamingResponse with SSE
-    When stream=False: Returns SearchFullResponse
+    If hybrid_alpha is unset and ONYX_SEARCH_UI_USES_OPENSEARCH_KEYWORD_SEARCH
+    is True, executes pure keyword search.
+
+    Returns:
+        StreamingResponse with SSE if stream=True, otherwise SearchFullResponse.
     """
     logger.debug(f"Received search query: {request.search_query}")
 
+    if request.hybrid_alpha is None and ONYX_SEARCH_UI_USES_OPENSEARCH_KEYWORD_SEARCH:
+        request.hybrid_alpha = 0.0
+
     # Non-streaming path
     if not request.stream:
         try:

backend/ee/onyx/server/seeding.py

@@ -178,7 +178,7 @@ def _seed_personas(db_session: Session, personas: list[PersonaUpsertRequest]) ->
                     system_prompt=persona.system_prompt,
                     task_prompt=persona.task_prompt,
                     datetime_aware=persona.datetime_aware,
-                    featured=persona.featured,
+                    is_featured=persona.is_featured,
                     commit=False,
                 )
             db_session.commit()

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

@@ -29,7 +29,6 @@ from onyx.configs.app_configs import OPENAI_DEFAULT_API_KEY
 from onyx.configs.app_configs import OPENROUTER_DEFAULT_API_KEY
 from onyx.configs.app_configs import VERTEXAI_DEFAULT_CREDENTIALS
 from onyx.configs.app_configs import VERTEXAI_DEFAULT_LOCATION
-from onyx.configs.constants import MilestoneRecordType
 from onyx.db.engine.sql_engine import get_session_with_shared_schema
 from onyx.db.engine.sql_engine import get_session_with_tenant
 from onyx.db.image_generation import create_default_image_gen_config_from_api_key
@@ -59,7 +58,6 @@ from onyx.server.manage.llm.models import LLMProviderUpsertRequest
 from onyx.server.manage.llm.models import ModelConfigurationUpsertRequest
 from onyx.setup import setup_onyx
 from onyx.utils.logger import setup_logger
-from onyx.utils.telemetry import mt_cloud_telemetry
 from shared_configs.configs import MULTI_TENANT
 from shared_configs.configs import POSTGRES_DEFAULT_SCHEMA
 from shared_configs.configs import TENANT_ID_PREFIX
@@ -71,7 +69,9 @@ logger = setup_logger()
 
 
 async def get_or_provision_tenant(
-    email: str, referral_source: str | None = None, request: Request | None = None
+    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.
@@ -99,6 +99,26 @@ async def get_or_provision_tenant(
         tenant_id = await get_available_tenant()
 
         if tenant_id:
+            # Run migrations to ensure the pre-provisioned tenant schema is current.
+            # Pool tenants may have been created before a new migration was deployed.
+            # Capture as a non-optional local so mypy can type the lambda correctly.
+            _tenant_id: str = tenant_id
+            loop = asyncio.get_running_loop()
+            try:
+                await loop.run_in_executor(
+                    None, lambda: run_alembic_migrations(_tenant_id)
+                )
+            except Exception:
+                # The tenant was already dequeued from the pool — roll it back so
+                # it doesn't end up orphaned (schema exists, but not assigned to anyone).
+                logger.exception(
+                    f"Migration failed for pre-provisioned tenant {_tenant_id}; rolling back"
+                )
+                try:
+                    await rollback_tenant_provisioning(_tenant_id)
+                except Exception:
+                    logger.exception(f"Failed to rollback orphaned tenant {_tenant_id}")
+                raise
             # If we have a pre-provisioned tenant, assign it to the user
             await assign_tenant_to_user(tenant_id, email, referral_source)
             logger.info(f"Assigned pre-provisioned tenant {tenant_id} to user {email}")
@@ -693,12 +713,6 @@ async def assign_tenant_to_user(
 
     try:
         add_users_to_tenant([email], tenant_id)
-
-        mt_cloud_telemetry(
-            tenant_id=tenant_id,
-            distinct_id=email,
-            event=MilestoneRecordType.TENANT_CREATED,
-        )
     except Exception:
         logger.exception(f"Failed to assign tenant {tenant_id} to user {email}")
         raise Exception("Failed to assign tenant to user")

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

@@ -4,6 +4,7 @@ from fastapi import HTTPException
 from sqlalchemy.exc import IntegrityError
 from sqlalchemy.orm import Session
 
+from ee.onyx.db.persona import update_persona_access
 from ee.onyx.db.user_group import add_users_to_user_group
 from ee.onyx.db.user_group import delete_user_group as db_delete_user_group
 from ee.onyx.db.user_group import fetch_user_group
@@ -11,13 +12,16 @@ from ee.onyx.db.user_group import fetch_user_groups
 from ee.onyx.db.user_group import fetch_user_groups_for_user
 from ee.onyx.db.user_group import insert_user_group
 from ee.onyx.db.user_group import prepare_user_group_for_deletion
+from ee.onyx.db.user_group import rename_user_group
 from ee.onyx.db.user_group import update_user_curator_relationship
 from ee.onyx.db.user_group import update_user_group
 from ee.onyx.server.user_group.models import AddUsersToUserGroupRequest
 from ee.onyx.server.user_group.models import MinimalUserGroupSnapshot
 from ee.onyx.server.user_group.models import SetCuratorRequest
+from ee.onyx.server.user_group.models import UpdateGroupAgentsRequest
 from ee.onyx.server.user_group.models import UserGroup
 from ee.onyx.server.user_group.models import UserGroupCreate
+from ee.onyx.server.user_group.models import UserGroupRename
 from ee.onyx.server.user_group.models import UserGroupUpdate
 from onyx.auth.users import current_admin_user
 from onyx.auth.users import current_curator_or_admin_user
@@ -27,6 +31,9 @@ from onyx.configs.constants import PUBLIC_API_TAGS
 from onyx.db.engine.sql_engine import get_session
 from onyx.db.models import User
 from onyx.db.models import UserRole
+from onyx.db.persona import get_persona_by_id
+from onyx.error_handling.error_codes import OnyxErrorCode
+from onyx.error_handling.exceptions import OnyxError
 from onyx.utils.logger import setup_logger
 
 logger = setup_logger()
@@ -87,6 +94,32 @@ def create_user_group(
     return UserGroup.from_model(db_user_group)
 
 
+@router.patch("/admin/user-group/rename")
+def rename_user_group_endpoint(
+    rename_request: UserGroupRename,
+    _: User = Depends(current_admin_user),
+    db_session: Session = Depends(get_session),
+) -> UserGroup:
+    try:
+        return UserGroup.from_model(
+            rename_user_group(
+                db_session=db_session,
+                user_group_id=rename_request.id,
+                new_name=rename_request.name,
+            )
+        )
+    except IntegrityError:
+        raise OnyxError(
+            OnyxErrorCode.DUPLICATE_RESOURCE,
+            f"User group with name '{rename_request.name}' already exists.",
+        )
+    except ValueError as e:
+        msg = str(e)
+        if "not found" in msg.lower():
+            raise OnyxError(OnyxErrorCode.NOT_FOUND, msg)
+        raise OnyxError(OnyxErrorCode.CONFLICT, msg)
+
+
 @router.patch("/admin/user-group/{user_group_id}")
 def patch_user_group(
     user_group_id: int,
@@ -161,3 +194,38 @@ def delete_user_group(
         user_group = fetch_user_group(db_session, user_group_id)
         if user_group:
             db_delete_user_group(db_session, user_group)
+
+
+@router.patch("/admin/user-group/{user_group_id}/agents")
+def update_group_agents(
+    user_group_id: int,
+    request: UpdateGroupAgentsRequest,
+    user: User = Depends(current_admin_user),
+    db_session: Session = Depends(get_session),
+) -> None:
+    for agent_id in request.added_agent_ids:
+        persona = get_persona_by_id(
+            persona_id=agent_id, user=user, db_session=db_session
+        )
+        current_group_ids = [g.id for g in persona.groups]
+        if user_group_id not in current_group_ids:
+            update_persona_access(
+                persona_id=agent_id,
+                creator_user_id=user.id,
+                db_session=db_session,
+                group_ids=current_group_ids + [user_group_id],
+            )
+
+    for agent_id in request.removed_agent_ids:
+        persona = get_persona_by_id(
+            persona_id=agent_id, user=user, db_session=db_session
+        )
+        current_group_ids = [g.id for g in persona.groups]
+        update_persona_access(
+            persona_id=agent_id,
+            creator_user_id=user.id,
+            db_session=db_session,
+            group_ids=[gid for gid in current_group_ids if gid != user_group_id],
+        )
+
+    db_session.commit()

backend/ee/onyx/server/user_group/models.py

@@ -104,6 +104,16 @@ class AddUsersToUserGroupRequest(BaseModel):
     user_ids: list[UUID]
 
 
+class UserGroupRename(BaseModel):
+    id: int
+    name: str
+
+
 class SetCuratorRequest(BaseModel):
     user_id: UUID
     is_curator: bool
+
+
+class UpdateGroupAgentsRequest(BaseModel):
+    added_agent_ids: list[int]
+    removed_agent_ids: list[int]

backend/ee/onyx/utils/posthog_client.py

@@ -9,6 +9,7 @@ from ee.onyx.configs.app_configs import POSTHOG_API_KEY
 from ee.onyx.configs.app_configs import POSTHOG_DEBUG_LOGS_ENABLED
 from ee.onyx.configs.app_configs import POSTHOG_HOST
 from onyx.utils.logger import setup_logger
+from shared_configs.configs import MULTI_TENANT
 
 logger = setup_logger()
 
@@ -18,12 +19,19 @@ def posthog_on_error(error: Any, items: Any) -> None:
     logger.error(f"PostHog error: {error}, items: {items}")
 
 
-posthog = Posthog(
-    project_api_key=POSTHOG_API_KEY,
-    host=POSTHOG_HOST,
-    debug=POSTHOG_DEBUG_LOGS_ENABLED,
-    on_error=posthog_on_error,
-)
+posthog: Posthog | None = None
+if POSTHOG_API_KEY:
+    posthog = Posthog(
+        project_api_key=POSTHOG_API_KEY,
+        host=POSTHOG_HOST,
+        debug=POSTHOG_DEBUG_LOGS_ENABLED,
+        on_error=posthog_on_error,
+    )
+elif MULTI_TENANT:
+    logger.warning(
+        "POSTHOG_API_KEY is not set but MULTI_TENANT is enabled — "
+        "PostHog telemetry and feature flags will be disabled"
+    )
 
 # For cross referencing between cloud and www Onyx sites
 # NOTE: These clients are separate because they are separate posthog projects.
@@ -60,7 +68,7 @@ def capture_and_sync_with_alternate_posthog(
         logger.error(f"Error capturing marketing posthog event: {e}")
 
     try:
-        if cloud_user_id := props.get("onyx_cloud_user_id"):
+        if posthog and (cloud_user_id := props.get("onyx_cloud_user_id")):
             cloud_props = props.copy()
             cloud_props.pop("onyx_cloud_user_id", None)
 
@@ -72,15 +80,45 @@ def capture_and_sync_with_alternate_posthog(
         logger.error(f"Error identifying cloud posthog user: {e}")
 
 
+def alias_user(distinct_id: str, anonymous_id: str) -> None:
+    """Link an anonymous distinct_id to an identified user, merging person profiles.
+
+    No-ops when the IDs match (e.g. returning users whose PostHog cookie
+    already contains their identified user ID).
+    """
+    if not posthog or anonymous_id == distinct_id:
+        return
+
+    try:
+        posthog.alias(previous_id=anonymous_id, distinct_id=distinct_id)
+        posthog.flush()
+    except Exception as e:
+        logger.error(f"Error aliasing PostHog user: {e}")
+
+
+def get_anon_id_from_request(request: Any) -> str | None:
+    """Extract the anonymous distinct_id from the app PostHog cookie on a request."""
+    if not POSTHOG_API_KEY:
+        return None
+
+    cookie_name = f"ph_{POSTHOG_API_KEY}_posthog"
+    if (cookie_value := request.cookies.get(cookie_name)) and (
+        parsed := parse_posthog_cookie(cookie_value)
+    ):
+        return parsed.get("distinct_id")
+
+    return None
+
+
 def get_marketing_posthog_cookie_name() -> str | None:
     if not MARKETING_POSTHOG_API_KEY:
         return None
     return f"onyx_custom_ph_{MARKETING_POSTHOG_API_KEY}_posthog"
 
 
-def parse_marketing_cookie(cookie_value: str) -> dict[str, Any] | None:
+def parse_posthog_cookie(cookie_value: str) -> dict[str, Any] | None:
     """
-    Parse the URL-encoded JSON marketing cookie.
+    Parse a URL-encoded JSON PostHog cookie
 
     Expected format (URL-encoded):
     {"distinct_id":"...", "featureFlags":{"landing_page_variant":"..."}, ...}
@@ -94,7 +132,7 @@ def parse_marketing_cookie(cookie_value: str) -> dict[str, Any] | None:
         cookie_data = json.loads(decoded_cookie)
 
         distinct_id = cookie_data.get("distinct_id")
-        if not distinct_id:
+        if not distinct_id or not isinstance(distinct_id, str):
             return None
 
         return cookie_data

backend/ee/onyx/utils/telemetry.py

@@ -1,3 +1,5 @@
+from typing import Any
+
 from ee.onyx.utils.posthog_client import posthog
 from onyx.utils.logger import setup_logger
 
@@ -5,12 +7,27 @@ logger = setup_logger()
 
 
 def event_telemetry(
-    distinct_id: str, event: str, properties: dict | None = None
+    distinct_id: str, event: str, properties: dict[str, Any] | None = None
 ) -> None:
     """Capture and send an event to PostHog, flushing immediately."""
+    if not posthog:
+        return
+
     logger.info(f"Capturing PostHog event: {distinct_id} {event} {properties}")
     try:
         posthog.capture(distinct_id, event, properties)
         posthog.flush()
     except Exception as e:
         logger.error(f"Error capturing PostHog event: {e}")
+
+
+def identify_user(distinct_id: str, properties: dict[str, Any] | None = None) -> None:
+    """Create/update a PostHog person profile, flushing immediately."""
+    if not posthog:
+        return
+
+    try:
+        posthog.identify(distinct_id, properties)
+        posthog.flush()
+    except Exception as e:
+        logger.error(f"Error identifying PostHog user: {e}")

backend/model_server/main.py

@@ -100,6 +100,7 @@ def get_model_app() -> FastAPI:
             dsn=SENTRY_DSN,
             integrations=[StarletteIntegration(), FastApiIntegration()],
             traces_sample_rate=0.1,
+            release=__version__,
         )
         logger.info("Sentry initialized")
     else:

backend/onyx/auth/users.py

@@ -19,6 +19,7 @@ from typing import Optional
 from typing import Protocol
 from typing import Tuple
 from typing import TypeVar
+from urllib.parse import urlparse
 
 import jwt
 from email_validator import EmailNotValidError
@@ -134,6 +135,9 @@ from onyx.redis.redis_pool import retrieve_ws_token_data
 from onyx.server.settings.store import load_settings
 from onyx.server.utils import BasicAuthenticationError
 from onyx.utils.logger import setup_logger
+from onyx.utils.telemetry import mt_cloud_alias
+from onyx.utils.telemetry import mt_cloud_get_anon_id
+from onyx.utils.telemetry import mt_cloud_identify
 from onyx.utils.telemetry import mt_cloud_telemetry
 from onyx.utils.telemetry import optional_telemetry
 from onyx.utils.telemetry import RecordType
@@ -249,18 +253,12 @@ def verify_email_is_invited(email: str) -> None:
     whitelist = get_invited_users()
 
     if not email:
-        raise HTTPException(
-            status_code=status.HTTP_400_BAD_REQUEST,
-            detail={"reason": "Email must be specified"},
-        )
+        raise OnyxError(OnyxErrorCode.INVALID_INPUT, "Email must be specified")
 
     try:
         email_info = validate_email(email, check_deliverability=False)
     except EmailUndeliverableError:
-        raise HTTPException(
-            status_code=status.HTTP_400_BAD_REQUEST,
-            detail={"reason": "Email is not valid"},
-        )
+        raise OnyxError(OnyxErrorCode.INVALID_INPUT, "Email is not valid")
 
     for email_whitelist in whitelist:
         try:
@@ -277,12 +275,9 @@ def verify_email_is_invited(email: str) -> None:
         if email_info.normalized.lower() == email_info_whitelist.normalized.lower():
             return
 
-    raise HTTPException(
-        status_code=status.HTTP_403_FORBIDDEN,
-        detail={
-            "code": REGISTER_INVITE_ONLY_CODE,
-            "reason": "This workspace is invite-only. Please ask your admin to invite you.",
-        },
+    raise OnyxError(
+        OnyxErrorCode.UNAUTHORIZED,
+        "This workspace is invite-only. Please ask your admin to invite you.",
     )
 
 
@@ -292,48 +287,47 @@ def verify_email_in_whitelist(email: str, tenant_id: str) -> None:
             verify_email_is_invited(email)
 
 
-def verify_email_domain(email: str) -> None:
+def verify_email_domain(email: str, *, is_registration: bool = False) -> None:
     if email.count("@") != 1:
-        raise HTTPException(
-            status_code=status.HTTP_400_BAD_REQUEST,
-            detail="Email is not valid",
-        )
+        raise OnyxError(OnyxErrorCode.INVALID_INPUT, "Email is not valid")
 
     local_part, domain = email.split("@")
     domain = domain.lower()
+    local_part = local_part.lower()
 
     if AUTH_TYPE == AuthType.CLOUD:
         # Normalize googlemail.com to gmail.com (they deliver to the same inbox)
         if domain == "googlemail.com":
-            raise HTTPException(
-                status_code=status.HTTP_400_BAD_REQUEST,
-                detail={"reason": "Please use @gmail.com instead of @googlemail.com."},
+            raise OnyxError(
+                OnyxErrorCode.INVALID_INPUT,
+                "Please use @gmail.com instead of @googlemail.com.",
+            )
+
+        # Only block dotted Gmail on new signups — existing users must still be
+        # able to sign in with the address they originally registered with.
+        if is_registration and domain == "gmail.com" and "." in local_part:
+            raise OnyxError(
+                OnyxErrorCode.INVALID_INPUT,
+                "Gmail addresses with '.' are not allowed. Please use your base email address.",
             )
 
         if "+" in local_part and domain != "onyx.app":
-            raise HTTPException(
-                status_code=status.HTTP_400_BAD_REQUEST,
-                detail={
-                    "reason": "Email addresses with '+' are not allowed. Please use your base email address."
-                },
+            raise OnyxError(
+                OnyxErrorCode.INVALID_INPUT,
+                "Email addresses with '+' are not allowed. Please use your base email address.",
             )
 
     # Check if email uses a disposable/temporary domain
     if is_disposable_email(email):
-        raise HTTPException(
-            status_code=status.HTTP_400_BAD_REQUEST,
-            detail={
-                "reason": "Disposable email addresses are not allowed. Please use a permanent email address."
-            },
+        raise OnyxError(
+            OnyxErrorCode.INVALID_INPUT,
+            "Disposable email addresses are not allowed. Please use a permanent email address.",
         )
 
     # Check domain whitelist if configured
     if VALID_EMAIL_DOMAINS:
         if domain not in VALID_EMAIL_DOMAINS:
-            raise HTTPException(
-                status_code=status.HTTP_400_BAD_REQUEST,
-                detail="Email domain is not valid",
-            )
+            raise OnyxError(OnyxErrorCode.INVALID_INPUT, "Email domain is not valid")
 
 
 def enforce_seat_limit(db_session: Session, seats_needed: int = 1) -> None:
@@ -349,7 +343,7 @@ def enforce_seat_limit(db_session: Session, seats_needed: int = 1) -> None:
     )(db_session, seats_needed=seats_needed)
 
     if result is not None and not result.available:
-        raise HTTPException(status_code=402, detail=result.error_message)
+        raise OnyxError(OnyxErrorCode.SEAT_LIMIT_EXCEEDED, result.error_message)
 
 
 class UserManager(UUIDIDMixin, BaseUserManager[User, uuid.UUID]):
@@ -402,10 +396,7 @@ class UserManager(UUIDIDMixin, BaseUserManager[User, uuid.UUID]):
                     captcha_token or "", expected_action="signup"
                 )
             except CaptchaVerificationError as e:
-                raise HTTPException(
-                    status_code=status.HTTP_400_BAD_REQUEST,
-                    detail={"reason": str(e)},
-                )
+                raise OnyxError(OnyxErrorCode.INVALID_INPUT, str(e))
 
         # We verify the password here to make sure it's valid before we proceed
         await self.validate_password(
@@ -415,13 +406,10 @@ class UserManager(UUIDIDMixin, BaseUserManager[User, uuid.UUID]):
         # Check for disposable emails BEFORE provisioning tenant
         # This prevents creating tenants for throwaway email addresses
         try:
-            verify_email_domain(user_create.email)
-        except HTTPException as e:
+            verify_email_domain(user_create.email, is_registration=True)
+        except OnyxError as e:
             # Log blocked disposable email attempts
-            if (
-                e.status_code == status.HTTP_400_BAD_REQUEST
-                and "Disposable email" in str(e.detail)
-            ):
+            if "Disposable email" in e.detail:
                 domain = (
                     user_create.email.split("@")[-1]
                     if "@" in user_create.email
@@ -565,9 +553,9 @@ class UserManager(UUIDIDMixin, BaseUserManager[User, uuid.UUID]):
         result = await db_session.execute(
             select(Persona.id)
             .where(
-                Persona.featured.is_(True),
+                Persona.is_featured.is_(True),
                 Persona.is_public.is_(True),
-                Persona.is_visible.is_(True),
+                Persona.is_listed.is_(True),
                 Persona.deleted.is_(False),
             )
             .order_by(
@@ -695,6 +683,8 @@ class UserManager(UUIDIDMixin, BaseUserManager[User, uuid.UUID]):
                         raise exceptions.UserNotExists()
 
                 except exceptions.UserNotExists:
+                    verify_email_domain(account_email, is_registration=True)
+
                     # Check seat availability before creating (single-tenant only)
                     with get_session_with_current_tenant() as sync_db:
                         enforce_seat_limit(sync_db)
@@ -792,6 +782,18 @@ class UserManager(UUIDIDMixin, BaseUserManager[User, uuid.UUID]):
         except Exception:
             logger.exception("Error deleting anonymous user cookie")
 
+        tenant_id = CURRENT_TENANT_ID_CONTEXTVAR.get()
+
+        # Link the anonymous PostHog session to the identified user so that
+        # pre-login session recordings and events merge into one person profile.
+        if anon_id := mt_cloud_get_anon_id(request):
+            mt_cloud_alias(distinct_id=str(user.id), anonymous_id=anon_id)
+
+        mt_cloud_identify(
+            distinct_id=str(user.id),
+            properties={"email": user.email, "tenant_id": tenant_id},
+        )
+
     async def on_after_register(
         self, user: User, request: Optional[Request] = None
     ) -> None:
@@ -810,12 +812,30 @@ class UserManager(UUIDIDMixin, BaseUserManager[User, uuid.UUID]):
             user_count = await get_user_count()
             logger.debug(f"Current tenant user count: {user_count}")
 
+            # Link the anonymous PostHog session to the identified user so
+            # that pre-signup session recordings merge into one person profile.
+            if anon_id := mt_cloud_get_anon_id(request):
+                mt_cloud_alias(distinct_id=str(user.id), anonymous_id=anon_id)
+
+            # Ensure a PostHog person profile exists for this user.
+            mt_cloud_identify(
+                distinct_id=str(user.id),
+                properties={"email": user.email, "tenant_id": tenant_id},
+            )
+
             mt_cloud_telemetry(
                 tenant_id=tenant_id,
-                distinct_id=user.email,
+                distinct_id=str(user.id),
                 event=MilestoneRecordType.USER_SIGNED_UP,
             )
 
+            if user_count == 1:
+                mt_cloud_telemetry(
+                    tenant_id=tenant_id,
+                    distinct_id=str(user.id),
+                    event=MilestoneRecordType.TENANT_CREATED,
+                )
+
         finally:
             CURRENT_TENANT_ID_CONTEXTVAR.reset(token)
 
@@ -825,9 +845,9 @@ class UserManager(UUIDIDMixin, BaseUserManager[User, uuid.UUID]):
             attribute="get_marketing_posthog_cookie_name",
             noop_return_value=None,
         )
-        parse_marketing_cookie = fetch_ee_implementation_or_noop(
+        parse_posthog_cookie = fetch_ee_implementation_or_noop(
             module="onyx.utils.posthog_client",
-            attribute="parse_marketing_cookie",
+            attribute="parse_posthog_cookie",
             noop_return_value=None,
         )
         capture_and_sync_with_alternate_posthog = fetch_ee_implementation_or_noop(
@@ -841,7 +861,7 @@ class UserManager(UUIDIDMixin, BaseUserManager[User, uuid.UUID]):
             and user_count is not None
             and (marketing_cookie_name := get_marketing_posthog_cookie_name())
             and (marketing_cookie_value := request.cookies.get(marketing_cookie_name))
-            and (parsed_cookie := parse_marketing_cookie(marketing_cookie_value))
+            and (parsed_cookie := parse_posthog_cookie(marketing_cookie_value))
         ):
             marketing_anonymous_id = parsed_cookie["distinct_id"]
 
@@ -1652,6 +1672,33 @@ async def _get_user_from_token_data(token_data: dict) -> User | None:
         return user
 
 
+_LOOPBACK_HOSTNAMES = frozenset({"localhost", "127.0.0.1", "::1"})
+
+
+def _is_same_origin(actual: str, expected: str) -> bool:
+    """Compare two origins for the WebSocket CSWSH check.
+
+    Scheme and hostname must match exactly.  Port must also match, except
+    when the hostname is a loopback address (localhost / 127.0.0.1 / ::1),
+    where port is ignored.  On loopback, all ports belong to the same
+    operator, so port differences carry no security significance — the
+    CSWSH threat is remote origins, not local ones.
+    """
+    a = urlparse(actual.rstrip("/"))
+    e = urlparse(expected.rstrip("/"))
+
+    if a.scheme != e.scheme or a.hostname != e.hostname:
+        return False
+
+    if a.hostname in _LOOPBACK_HOSTNAMES:
+        return True
+
+    actual_port = a.port or (443 if a.scheme == "https" else 80)
+    expected_port = e.port or (443 if e.scheme == "https" else 80)
+
+    return actual_port == expected_port
+
+
 async def current_user_from_websocket(
     websocket: WebSocket,
     token: str = Query(..., description="WebSocket authentication token"),
@@ -1671,19 +1718,15 @@ async def current_user_from_websocket(
 
     This applies the same auth checks as current_user() for HTTP endpoints.
     """
-    # Check Origin header to prevent Cross-Site WebSocket Hijacking (CSWSH)
-    # Browsers always send Origin on WebSocket connections
+    # Check Origin header to prevent Cross-Site WebSocket Hijacking (CSWSH).
+    # Browsers always send Origin on WebSocket connections.
     origin = websocket.headers.get("origin")
-    expected_origin = WEB_DOMAIN.rstrip("/")
     if not origin:
         logger.warning("WS auth: missing Origin header")
         raise BasicAuthenticationError(detail="Access denied. Missing origin.")
 
-    actual_origin = origin.rstrip("/")
-    if actual_origin != expected_origin:
-        logger.warning(
-            f"WS auth: origin mismatch. Expected {expected_origin}, got {actual_origin}"
-        )
+    if not _is_same_origin(origin, WEB_DOMAIN):
+        logger.warning(f"WS auth: origin mismatch. Expected {WEB_DOMAIN}, got {origin}")
         raise BasicAuthenticationError(detail="Access denied. Invalid origin.")
 
     # Validate WS token in Redis (single-use, deleted after retrieval)

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

@@ -20,6 +20,7 @@ from sentry_sdk.integrations.celery import CeleryIntegration
 from sqlalchemy import text
 from sqlalchemy.orm import Session
 
+from onyx import __version__
 from onyx.background.celery.apps.task_formatters import CeleryTaskColoredFormatter
 from onyx.background.celery.apps.task_formatters import CeleryTaskPlainFormatter
 from onyx.background.celery.celery_utils import celery_is_worker_primary
@@ -65,6 +66,7 @@ if SENTRY_DSN:
         dsn=SENTRY_DSN,
         integrations=[CeleryIntegration()],
         traces_sample_rate=0.1,
+        release=__version__,
     )
     logger.info("Sentry initialized")
 else:
@@ -515,7 +517,8 @@ def reset_tenant_id(
 
 
 def wait_for_vespa_or_shutdown(
-    sender: Any, **kwargs: Any  # noqa: ARG001
+    sender: Any,  # noqa: ARG001
+    **kwargs: Any,  # noqa: ARG001
 ) -> None:  # noqa: ARG001
     """Waits for Vespa to become ready subject to a timeout.
     Raises WorkerShutdown if the timeout is reached."""

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

@@ -13,6 +13,14 @@ from celery.signals import worker_shutdown
 import onyx.background.celery.apps.app_base as app_base
 from onyx.configs.constants import POSTGRES_CELERY_WORKER_DOCFETCHING_APP_NAME
 from onyx.db.engine.sql_engine import SqlEngine
+from onyx.server.metrics.celery_task_metrics import on_celery_task_postrun
+from onyx.server.metrics.celery_task_metrics import on_celery_task_prerun
+from onyx.server.metrics.celery_task_metrics import on_celery_task_rejected
+from onyx.server.metrics.celery_task_metrics import on_celery_task_retry
+from onyx.server.metrics.celery_task_metrics import on_celery_task_revoked
+from onyx.server.metrics.indexing_task_metrics import on_indexing_task_postrun
+from onyx.server.metrics.indexing_task_metrics import on_indexing_task_prerun
+from onyx.server.metrics.metrics_server import start_metrics_server
 from onyx.utils.logger import setup_logger
 from shared_configs.configs import MULTI_TENANT
 
@@ -34,6 +42,8 @@ def on_task_prerun(
     **kwds: Any,
 ) -> None:
     app_base.on_task_prerun(sender, task_id, task, args, kwargs, **kwds)
+    on_celery_task_prerun(task_id, task)
+    on_indexing_task_prerun(task_id, task, kwargs)
 
 
 @signals.task_postrun.connect
@@ -48,6 +58,36 @@ def on_task_postrun(
     **kwds: Any,
 ) -> None:
     app_base.on_task_postrun(sender, task_id, task, args, kwargs, retval, state, **kwds)
+    on_celery_task_postrun(task_id, task, state)
+    on_indexing_task_postrun(task_id, task, kwargs, state)
+
+
+@signals.task_retry.connect
+def on_task_retry(sender: Any | None = None, **kwargs: Any) -> None:  # noqa: ARG001
+    # task_retry signal doesn't pass task_id in kwargs; get it from
+    # the sender (the task instance) via sender.request.id.
+    task_id = getattr(getattr(sender, "request", None), "id", None)
+    on_celery_task_retry(task_id, sender)
+
+
+@signals.task_revoked.connect
+def on_task_revoked(sender: Any | None = None, **kwargs: Any) -> None:
+    task_name = getattr(sender, "name", None) or str(sender)
+    on_celery_task_revoked(kwargs.get("task_id"), task_name)
+
+
+@signals.task_rejected.connect
+def on_task_rejected(sender: Any | None = None, **kwargs: Any) -> None:  # noqa: ARG001
+    # task_rejected sends the Consumer as sender, not the task instance.
+    # The task name must be extracted from the Celery message headers.
+    message = kwargs.get("message")
+    task_name: str | None = None
+    if message is not None:
+        headers = getattr(message, "headers", None) or {}
+        task_name = headers.get("task")
+    if task_name is None:
+        task_name = "unknown"
+    on_celery_task_rejected(None, task_name)
 
 
 @celeryd_init.connect
@@ -76,6 +116,7 @@ def on_worker_init(sender: Worker, **kwargs: Any) -> None:
 
 @worker_ready.connect
 def on_worker_ready(sender: Any, **kwargs: Any) -> None:
+    start_metrics_server("docfetching")
     app_base.on_worker_ready(sender, **kwargs)
 
 

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

@@ -14,6 +14,14 @@ from celery.signals import worker_shutdown
 import onyx.background.celery.apps.app_base as app_base
 from onyx.configs.constants import POSTGRES_CELERY_WORKER_DOCPROCESSING_APP_NAME
 from onyx.db.engine.sql_engine import SqlEngine
+from onyx.server.metrics.celery_task_metrics import on_celery_task_postrun
+from onyx.server.metrics.celery_task_metrics import on_celery_task_prerun
+from onyx.server.metrics.celery_task_metrics import on_celery_task_rejected
+from onyx.server.metrics.celery_task_metrics import on_celery_task_retry
+from onyx.server.metrics.celery_task_metrics import on_celery_task_revoked
+from onyx.server.metrics.indexing_task_metrics import on_indexing_task_postrun
+from onyx.server.metrics.indexing_task_metrics import on_indexing_task_prerun
+from onyx.server.metrics.metrics_server import start_metrics_server
 from onyx.utils.logger import setup_logger
 from shared_configs.configs import MULTI_TENANT
 
@@ -35,6 +43,8 @@ def on_task_prerun(
     **kwds: Any,
 ) -> None:
     app_base.on_task_prerun(sender, task_id, task, args, kwargs, **kwds)
+    on_celery_task_prerun(task_id, task)
+    on_indexing_task_prerun(task_id, task, kwargs)
 
 
 @signals.task_postrun.connect
@@ -49,6 +59,36 @@ def on_task_postrun(
     **kwds: Any,
 ) -> None:
     app_base.on_task_postrun(sender, task_id, task, args, kwargs, retval, state, **kwds)
+    on_celery_task_postrun(task_id, task, state)
+    on_indexing_task_postrun(task_id, task, kwargs, state)
+
+
+@signals.task_retry.connect
+def on_task_retry(sender: Any | None = None, **kwargs: Any) -> None:  # noqa: ARG001
+    # task_retry signal doesn't pass task_id in kwargs; get it from
+    # the sender (the task instance) via sender.request.id.
+    task_id = getattr(getattr(sender, "request", None), "id", None)
+    on_celery_task_retry(task_id, sender)
+
+
+@signals.task_revoked.connect
+def on_task_revoked(sender: Any | None = None, **kwargs: Any) -> None:
+    task_name = getattr(sender, "name", None) or str(sender)
+    on_celery_task_revoked(kwargs.get("task_id"), task_name)
+
+
+@signals.task_rejected.connect
+def on_task_rejected(sender: Any | None = None, **kwargs: Any) -> None:  # noqa: ARG001
+    # task_rejected sends the Consumer as sender, not the task instance.
+    # The task name must be extracted from the Celery message headers.
+    message = kwargs.get("message")
+    task_name: str | None = None
+    if message is not None:
+        headers = getattr(message, "headers", None) or {}
+        task_name = headers.get("task")
+    if task_name is None:
+        task_name = "unknown"
+    on_celery_task_rejected(None, task_name)
 
 
 @celeryd_init.connect
@@ -82,6 +122,7 @@ def on_worker_init(sender: Worker, **kwargs: Any) -> None:
 
 @worker_ready.connect
 def on_worker_ready(sender: Any, **kwargs: Any) -> None:
+    start_metrics_server("docprocessing")
     app_base.on_worker_ready(sender, **kwargs)
 
 
@@ -90,6 +131,12 @@ def on_worker_shutdown(sender: Any, **kwargs: Any) -> None:
     app_base.on_worker_shutdown(sender, **kwargs)
 
 
+# Note: worker_process_init only fires in prefork pool mode. Docprocessing uses
+# worker_pool="threads" (see configs/docprocessing.py), so this handler is
+# effectively a no-op in normal operation. It remains as a safety net in case
+# the pool type is ever changed to prefork. Prometheus metrics are safe in
+# thread-pool mode since all threads share the same process memory and can
+# update the same Counter/Gauge/Histogram objects directly.
 @worker_process_init.connect
 def init_worker(**kwargs: Any) -> None:  # noqa: ARG001
     SqlEngine.reset_engine()

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

@@ -54,8 +54,14 @@ def on_celeryd_init(sender: Any = None, conf: Any = None, **kwargs: Any) -> None
     app_base.on_celeryd_init(sender, conf, **kwargs)
 
 
+# Set by on_worker_init so on_worker_ready knows whether to start the server.
+_prometheus_collectors_ok: bool = False
+
+
 @worker_init.connect
 def on_worker_init(sender: Any, **kwargs: Any) -> None:
+    global _prometheus_collectors_ok
+
     logger.info("worker_init signal received.")
     logger.info(f"Multiprocessing start method: {multiprocessing.get_start_method()}")
 
@@ -65,6 +71,8 @@ def on_worker_init(sender: Any, **kwargs: Any) -> None:
     app_base.wait_for_redis(sender, **kwargs)
     app_base.wait_for_db(sender, **kwargs)
 
+    _prometheus_collectors_ok = _setup_prometheus_collectors(sender)
+
     # Less startup checks in multi-tenant case
     if MULTI_TENANT:
         return
@@ -72,8 +80,37 @@ def on_worker_init(sender: Any, **kwargs: Any) -> None:
     app_base.on_secondary_worker_init(sender, **kwargs)
 
 
+def _setup_prometheus_collectors(sender: Any) -> bool:
+    """Register Prometheus collectors that need Redis/DB access.
+
+    Passes the Celery app so the queue depth collector can obtain a fresh
+    broker Redis client on each scrape (rather than holding a stale reference).
+
+    Returns True if registration succeeded, False otherwise.
+    """
+    try:
+        from onyx.server.metrics.indexing_pipeline_setup import (
+            setup_indexing_pipeline_metrics,
+        )
+
+        setup_indexing_pipeline_metrics(sender.app)
+        logger.info("Prometheus indexing pipeline collectors registered")
+        return True
+    except Exception:
+        logger.exception("Failed to register Prometheus indexing pipeline collectors")
+        return False
+
+
 @worker_ready.connect
 def on_worker_ready(sender: Any, **kwargs: Any) -> None:
+    if _prometheus_collectors_ok:
+        from onyx.server.metrics.metrics_server import start_metrics_server
+
+        start_metrics_server("monitoring")
+    else:
+        logger.warning(
+            "Skipping Prometheus metrics server — collector registration failed"
+        )
     app_base.on_worker_ready(sender, **kwargs)
 
 

backend/onyx/background/celery/celery_redis.py

@@ -1,5 +1,6 @@
 # These are helper objects for tracking the keys we need to write in redis
 import json
+import threading
 from typing import Any
 from typing import cast
 
@@ -7,7 +8,59 @@ from celery import Celery
 from redis import Redis
 
 from onyx.background.celery.configs.base import CELERY_SEPARATOR
+from onyx.configs.app_configs import REDIS_HEALTH_CHECK_INTERVAL
 from onyx.configs.constants import OnyxCeleryPriority
+from onyx.configs.constants import REDIS_SOCKET_KEEPALIVE_OPTIONS
+
+
+_broker_client: Redis | None = None
+_broker_url: str | None = None
+_broker_client_lock = threading.Lock()
+
+
+def celery_get_broker_client(app: Celery) -> Redis:
+    """Return a shared Redis client connected to the Celery broker DB.
+
+    Uses a module-level singleton so all tasks on a worker share one
+    connection instead of creating a new one per call. The client
+    connects directly to the broker Redis DB (parsed from the broker URL).
+
+    Thread-safe via lock — safe for use in Celery thread-pool workers.
+
+    Usage:
+        r_celery = celery_get_broker_client(self.app)
+        length = celery_get_queue_length(queue, r_celery)
+    """
+    global _broker_client, _broker_url
+    with _broker_client_lock:
+        url = app.conf.broker_url
+        if _broker_client is not None and _broker_url == url:
+            try:
+                _broker_client.ping()
+                return _broker_client
+            except Exception:
+                try:
+                    _broker_client.close()
+                except Exception:
+                    pass
+                _broker_client = None
+        elif _broker_client is not None:
+            try:
+                _broker_client.close()
+            except Exception:
+                pass
+            _broker_client = None
+
+        _broker_url = url
+        _broker_client = Redis.from_url(
+            url,
+            decode_responses=False,
+            health_check_interval=REDIS_HEALTH_CHECK_INTERVAL,
+            socket_keepalive=True,
+            socket_keepalive_options=REDIS_SOCKET_KEEPALIVE_OPTIONS,
+            retry_on_timeout=True,
+        )
+        return _broker_client
 
 
 def celery_get_unacked_length(r: Redis) -> int:

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

@@ -14,6 +14,7 @@ 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_broker_client
 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
@@ -132,7 +133,6 @@ def revoke_tasks_blocking_deletion(
 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,
@@ -149,6 +149,7 @@ def check_for_connector_deletion_task(self: Task, *, tenant_id: str) -> bool | N
         if not r.exists(OnyxRedisSignals.BLOCK_VALIDATE_CONNECTOR_DELETION_FENCES):
             # clear fences that don't have associated celery tasks in progress
             try:
+                r_celery = celery_get_broker_client(self.app)
                 validate_connector_deletion_fences(
                     tenant_id, r, r_replica, r_celery, lock_beat
                 )

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

@@ -9,6 +9,7 @@ from celery import Celery
 from celery import shared_task
 from celery import Task
 
+from onyx import __version__
 from onyx.background.celery.apps.app_base import task_logger
 from onyx.background.celery.memory_monitoring import emit_process_memory
 from onyx.background.celery.tasks.docprocessing.heartbeat import start_heartbeat
@@ -137,6 +138,7 @@ def _docfetching_task(
         sentry_sdk.init(
             dsn=SENTRY_DSN,
             traces_sample_rate=0.1,
+            release=__version__,
         )
         logger.info("Sentry initialized")
     else:

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

@@ -22,6 +22,7 @@ from sqlalchemy.orm import Session
 
 from onyx.background.celery.apps.app_base import task_logger
 from onyx.background.celery.celery_redis import celery_find_task
+from onyx.background.celery.celery_redis import celery_get_broker_client
 from onyx.background.celery.celery_redis import celery_get_unacked_task_ids
 from onyx.background.celery.celery_utils import httpx_init_vespa_pool
 from onyx.background.celery.memory_monitoring import emit_process_memory
@@ -318,6 +319,11 @@ def monitor_indexing_attempt_progress(
     )
 
     current_db_time = get_db_current_time(db_session)
+    total_batches: int | str = (
+        coordination_status.total_batches
+        if coordination_status.total_batches is not None
+        else "?"
+    )
     if coordination_status.found:
         task_logger.info(
             f"Indexing attempt progress: "
@@ -325,7 +331,7 @@ def monitor_indexing_attempt_progress(
             f"cc_pair={attempt.connector_credential_pair_id} "
             f"search_settings={attempt.search_settings_id} "
             f"completed_batches={coordination_status.completed_batches} "
-            f"total_batches={coordination_status.total_batches or '?'} "
+            f"total_batches={total_batches} "
             f"total_docs={coordination_status.total_docs} "
             f"total_failures={coordination_status.total_failures}"
             f"elapsed={(current_db_time - attempt.time_created).seconds}"
@@ -409,7 +415,7 @@ def check_indexing_completion(
     logger.info(
         f"Indexing status: "
         f"indexing_completed={indexing_completed} "
-        f"batches_processed={batches_processed}/{batches_total or '?'} "
+        f"batches_processed={batches_processed}/{batches_total if batches_total is not None else '?'} "
         f"total_docs={coordination_status.total_docs} "
         f"total_chunks={coordination_status.total_chunks} "
         f"total_failures={coordination_status.total_failures}"
@@ -449,7 +455,7 @@ def check_indexing_completion(
             ):
                 # Check if the task exists in the celery queue
                 # This handles the case where Redis dies after task creation but before task execution
-                redis_celery = task.app.broker_connection().channel().client  # type: ignore
+                redis_celery = celery_get_broker_client(task.app)
                 task_exists = celery_find_task(
                     attempt.celery_task_id,
                     OnyxCeleryQueues.CONNECTOR_DOC_FETCHING,

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

@@ -29,6 +29,8 @@ from onyx.configs.constants import OnyxCeleryPriority
 from onyx.configs.constants import OnyxCeleryQueues
 from onyx.configs.constants import OnyxCeleryTask
 from onyx.configs.constants import OnyxRedisLocks
+from onyx.connectors.factory import ConnectorMissingException
+from onyx.connectors.factory import identify_connector_class
 from onyx.connectors.factory import instantiate_connector
 from onyx.connectors.interfaces import HierarchyConnector
 from onyx.connectors.models import HierarchyNode as PydanticHierarchyNode
@@ -55,6 +57,26 @@ logger = setup_logger()
 HIERARCHY_FETCH_INTERVAL_SECONDS = 24 * 60 * 60
 
 
+def _connector_supports_hierarchy_fetching(
+    cc_pair: ConnectorCredentialPair,
+) -> bool:
+    """Return True only for connectors whose class implements HierarchyConnector."""
+    try:
+        connector_class = identify_connector_class(
+            cc_pair.connector.source,
+        )
+    except ConnectorMissingException as e:
+        task_logger.warning(
+            "Skipping hierarchy fetching enqueue for source=%s input_type=%s: %s",
+            cc_pair.connector.source,
+            cc_pair.connector.input_type,
+            str(e),
+        )
+        return False
+
+    return issubclass(connector_class, HierarchyConnector)
+
+
 def _is_hierarchy_fetching_due(cc_pair: ConnectorCredentialPair) -> bool:
     """Returns boolean indicating if hierarchy fetching is due for this connector.
 
@@ -186,7 +208,10 @@ def check_for_hierarchy_fetching(self: Task, *, tenant_id: str) -> int | None:
                     cc_pair_id=cc_pair_id,
                 )
 
-                if not cc_pair or not _is_hierarchy_fetching_due(cc_pair):
+                if not cc_pair or not _connector_supports_hierarchy_fetching(cc_pair):
+                    continue
+
+                if not _is_hierarchy_fetching_due(cc_pair):
                     continue
 
                 task_id = _try_creating_hierarchy_fetching_task(

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

@@ -1,6 +1,5 @@
 import json
 import time
-from collections.abc import Callable
 from datetime import timedelta
 from itertools import islice
 from typing import Any
@@ -19,6 +18,7 @@ from sqlalchemy import text
 from sqlalchemy.orm import Session
 
 from onyx.background.celery.apps.app_base import task_logger
+from onyx.background.celery.celery_redis import celery_get_broker_client
 from onyx.background.celery.celery_redis import celery_get_queue_length
 from onyx.background.celery.celery_redis import celery_get_unacked_task_ids
 from onyx.background.celery.memory_monitoring import emit_process_memory
@@ -698,31 +698,27 @@ def monitor_background_processes(self: Task, *, tenant_id: str) -> None:
         return None
 
     try:
-        # Get Redis client for Celery broker
-        redis_celery = self.app.broker_connection().channel().client  # type: ignore
         redis_std = get_redis_client()
 
-        # Define metric collection functions and their dependencies
-        metric_functions: list[Callable[[], list[Metric]]] = [
-            lambda: _collect_queue_metrics(redis_celery),
-            lambda: _collect_connector_metrics(db_session, redis_std),
-            lambda: _collect_sync_metrics(db_session, redis_std),
-        ]
+        # Collect queue metrics with broker connection
+        r_celery = celery_get_broker_client(self.app)
+        queue_metrics = _collect_queue_metrics(r_celery)
 
-        # Collect and log each metric
+        # Collect remaining metrics (no broker connection needed)
         with get_session_with_current_tenant() as db_session:
-            for metric_fn in metric_functions:
-                metrics = metric_fn()
-                for metric in metrics:
-                    # double check to make sure we aren't double-emitting metrics
-                    if metric.key is None or not _has_metric_been_emitted(
-                        redis_std, metric.key
-                    ):
-                        metric.log()
-                        metric.emit(tenant_id)
+            all_metrics: list[Metric] = queue_metrics
+            all_metrics.extend(_collect_connector_metrics(db_session, redis_std))
+            all_metrics.extend(_collect_sync_metrics(db_session, redis_std))
 
-                    if metric.key is not None:
-                        _mark_metric_as_emitted(redis_std, metric.key)
+            for metric in all_metrics:
+                if metric.key is None or not _has_metric_been_emitted(
+                    redis_std, metric.key
+                ):
+                    metric.log()
+                    metric.emit(tenant_id)
+
+                if metric.key is not None:
+                    _mark_metric_as_emitted(redis_std, metric.key)
 
         task_logger.info("Successfully collected background metrics")
     except SoftTimeLimitExceeded:
@@ -890,7 +886,7 @@ def monitor_celery_queues_helper(
 ) -> None:
     """A task to monitor all celery queue lengths."""
 
-    r_celery = task.app.broker_connection().channel().client  # type: ignore
+    r_celery = celery_get_broker_client(task.app)
     n_celery = celery_get_queue_length(OnyxCeleryQueues.PRIMARY, r_celery)
     n_docfetching = celery_get_queue_length(
         OnyxCeleryQueues.CONNECTOR_DOC_FETCHING, r_celery
@@ -1080,7 +1076,7 @@ def cloud_monitor_celery_pidbox(
     num_deleted = 0
 
     MAX_PIDBOX_IDLE = 24 * 3600  # 1 day in seconds
-    r_celery: Redis = self.app.broker_connection().channel().client  # type: ignore
+    r_celery = celery_get_broker_client(self.app)
     for key in r_celery.scan_iter("*.reply.celery.pidbox"):
         key_bytes = cast(bytes, key)
         key_str = key_bytes.decode("utf-8")

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

@@ -17,6 +17,7 @@ from sqlalchemy.orm import Session
 
 from onyx.background.celery.apps.app_base import task_logger
 from onyx.background.celery.celery_redis import celery_find_task
+from onyx.background.celery.celery_redis import celery_get_broker_client
 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
@@ -203,7 +204,6 @@ def _is_pruning_due(cc_pair: ConnectorCredentialPair) -> bool:
 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(
         OnyxRedisLocks.CHECK_PRUNE_BEAT_LOCK,
@@ -261,6 +261,7 @@ def check_for_pruning(self: Task, *, tenant_id: str) -> bool | None:
             # tasks can be in the queue in redis, in reserved tasks (prefetched by the worker),
             # or be currently executing
             try:
+                r_celery = celery_get_broker_client(self.app)
                 validate_pruning_fences(tenant_id, r, r_replica, r_celery, lock_beat)
             except Exception:
                 task_logger.exception("Exception while validating pruning fences")

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

@@ -16,6 +16,7 @@ from sqlalchemy.orm import Session
 
 from onyx.access.access import build_access_for_user_files
 from onyx.background.celery.apps.app_base import task_logger
+from onyx.background.celery.celery_redis import celery_get_broker_client
 from onyx.background.celery.celery_redis import celery_get_queue_length
 from onyx.background.celery.celery_utils import httpx_init_vespa_pool
 from onyx.background.celery.tasks.shared.RetryDocumentIndex import RetryDocumentIndex
@@ -24,6 +25,7 @@ from onyx.configs.app_configs import MANAGED_VESPA
 from onyx.configs.app_configs import VESPA_CLOUD_CERT_PATH
 from onyx.configs.app_configs import VESPA_CLOUD_KEY_PATH
 from onyx.configs.constants import CELERY_GENERIC_BEAT_LOCK_TIMEOUT
+from onyx.configs.constants import CELERY_USER_FILE_DELETE_TASK_EXPIRES
 from onyx.configs.constants import CELERY_USER_FILE_PROCESSING_LOCK_TIMEOUT
 from onyx.configs.constants import CELERY_USER_FILE_PROCESSING_TASK_EXPIRES
 from onyx.configs.constants import CELERY_USER_FILE_PROJECT_SYNC_LOCK_TIMEOUT
@@ -33,6 +35,7 @@ from onyx.configs.constants import OnyxCeleryPriority
 from onyx.configs.constants import OnyxCeleryQueues
 from onyx.configs.constants import OnyxCeleryTask
 from onyx.configs.constants import OnyxRedisLocks
+from onyx.configs.constants import USER_FILE_DELETE_MAX_QUEUE_DEPTH
 from onyx.configs.constants import USER_FILE_PROCESSING_MAX_QUEUE_DEPTH
 from onyx.configs.constants import USER_FILE_PROJECT_SYNC_MAX_QUEUE_DEPTH
 from onyx.connectors.file.connector import LocalFileConnector
@@ -91,8 +94,19 @@ def _user_file_delete_lock_key(user_file_id: str | UUID) -> str:
     return f"{OnyxRedisLocks.USER_FILE_DELETE_LOCK_PREFIX}:{user_file_id}"
 
 
+def _user_file_delete_queued_key(user_file_id: str | UUID) -> str:
+    """Key that exists while a delete_single_user_file task is sitting in the queue.
+
+    The beat generator sets this with a TTL equal to CELERY_USER_FILE_DELETE_TASK_EXPIRES
+    before enqueuing and the worker deletes it as its first action.  This prevents
+    the beat from adding duplicate tasks for files that already have a live task
+    in flight.
+    """
+    return f"{OnyxRedisLocks.USER_FILE_DELETE_QUEUED_PREFIX}:{user_file_id}"
+
+
 def get_user_file_project_sync_queue_depth(celery_app: Celery) -> int:
-    redis_celery: Redis = celery_app.broker_connection().channel().client  # type: ignore
+    redis_celery = celery_get_broker_client(celery_app)
     return celery_get_queue_length(
         OnyxCeleryQueues.USER_FILE_PROJECT_SYNC, redis_celery
     )
@@ -225,7 +239,7 @@ def check_user_file_processing(self: Task, *, tenant_id: str) -> None:
     skipped_guard = 0
     try:
         # --- Protection 1: queue depth backpressure ---
-        r_celery = self.app.broker_connection().channel().client  # type: ignore
+        r_celery = celery_get_broker_client(self.app)
         queue_len = celery_get_queue_length(
             OnyxCeleryQueues.USER_FILE_PROCESSING, r_celery
         )
@@ -546,7 +560,23 @@ def process_single_user_file(
     ignore_result=True,
 )
 def check_for_user_file_delete(self: Task, *, tenant_id: str) -> None:
-    """Scan for user files with DELETING status and enqueue per-file tasks."""
+    """Scan for user files with DELETING status and enqueue per-file tasks.
+
+    Three mechanisms prevent queue runaway (mirrors check_user_file_processing):
+
+    1. **Queue depth backpressure** – if the broker queue already has more than
+       USER_FILE_DELETE_MAX_QUEUE_DEPTH items we skip this beat cycle entirely.
+
+    2. **Per-file queued guard** – before enqueuing a task we set a short-lived
+       Redis key (TTL = CELERY_USER_FILE_DELETE_TASK_EXPIRES).  If that key
+       already exists the file already has a live task in the queue, so we skip
+       it.  The worker deletes the key the moment it picks up the task so the
+       next beat cycle can re-enqueue if the file is still DELETING.
+
+    3. **Task expiry** – every enqueued task carries an `expires` value equal to
+       CELERY_USER_FILE_DELETE_TASK_EXPIRES.  If a task is still sitting in
+       the queue after that deadline, Celery discards it without touching the DB.
+    """
     task_logger.info("check_for_user_file_delete - Starting")
     redis_client = get_redis_client(tenant_id=tenant_id)
     lock: RedisLock = redis_client.lock(
@@ -555,8 +585,23 @@ def check_for_user_file_delete(self: Task, *, tenant_id: str) -> None:
     )
     if not lock.acquire(blocking=False):
         return None
+
     enqueued = 0
+    skipped_guard = 0
     try:
+        # --- Protection 1: queue depth backpressure ---
+        # NOTE: must use the broker's Redis client (not redis_client) because
+        # Celery queues live on a separate Redis DB with CELERY_SEPARATOR keys.
+        r_celery = celery_get_broker_client(self.app)
+        queue_len = celery_get_queue_length(OnyxCeleryQueues.USER_FILE_DELETE, r_celery)
+        if queue_len > USER_FILE_DELETE_MAX_QUEUE_DEPTH:
+            task_logger.warning(
+                f"check_for_user_file_delete - Queue depth {queue_len} exceeds "
+                f"{USER_FILE_DELETE_MAX_QUEUE_DEPTH}, skipping enqueue for "
+                f"tenant={tenant_id}"
+            )
+            return None
+
         with get_session_with_current_tenant() as db_session:
             user_file_ids = (
                 db_session.execute(
@@ -568,23 +613,40 @@ def check_for_user_file_delete(self: Task, *, tenant_id: str) -> None:
                 .all()
             )
             for user_file_id in user_file_ids:
-                self.app.send_task(
-                    OnyxCeleryTask.DELETE_SINGLE_USER_FILE,
-                    kwargs={"user_file_id": str(user_file_id), "tenant_id": tenant_id},
-                    queue=OnyxCeleryQueues.USER_FILE_DELETE,
-                    priority=OnyxCeleryPriority.HIGH,
+                # --- Protection 2: per-file queued guard ---
+                queued_key = _user_file_delete_queued_key(user_file_id)
+                guard_set = redis_client.set(
+                    queued_key,
+                    1,
+                    ex=CELERY_USER_FILE_DELETE_TASK_EXPIRES,
+                    nx=True,
                 )
+                if not guard_set:
+                    skipped_guard += 1
+                    continue
+
+                # --- Protection 3: task expiry ---
+                try:
+                    self.app.send_task(
+                        OnyxCeleryTask.DELETE_SINGLE_USER_FILE,
+                        kwargs={
+                            "user_file_id": str(user_file_id),
+                            "tenant_id": tenant_id,
+                        },
+                        queue=OnyxCeleryQueues.USER_FILE_DELETE,
+                        priority=OnyxCeleryPriority.HIGH,
+                        expires=CELERY_USER_FILE_DELETE_TASK_EXPIRES,
+                    )
+                except Exception:
+                    redis_client.delete(queued_key)
+                    raise
                 enqueued += 1
-    except Exception as e:
-        task_logger.exception(
-            f"check_for_user_file_delete - Error enqueuing deletes - {e.__class__.__name__}"
-        )
-        return None
     finally:
         if lock.owned():
             lock.release()
+
     task_logger.info(
-        f"check_for_user_file_delete - Enqueued {enqueued} tasks for tenant={tenant_id}"
+        f"check_for_user_file_delete - Enqueued {enqueued} tasks, skipped_guard={skipped_guard} for tenant={tenant_id}"
     )
     return None
 
@@ -602,6 +664,9 @@ def delete_user_file_impl(
     file_lock: RedisLock | None = None
     if redis_locking:
         redis_client = get_redis_client(tenant_id=tenant_id)
+        # Clear the queued guard so the beat can re-enqueue if deletion fails
+        # and the file remains in DELETING status.
+        redis_client.delete(_user_file_delete_queued_key(user_file_id))
         file_lock = redis_client.lock(
             _user_file_delete_lock_key(user_file_id),
             timeout=CELERY_GENERIC_BEAT_LOCK_TIMEOUT,

backend/onyx/cache/postgres_backend.py

@@ -297,7 +297,9 @@ class PostgresCacheBackend(CacheBackend):
 
     def _lock_id_for(self, name: str) -> int:
         """Map *name* to a 64-bit signed int for ``pg_advisory_lock``."""
-        h = hashlib.md5(f"{self._tenant_id}:{name}".encode()).digest()
+        h = hashlib.md5(
+            f"{self._tenant_id}:{name}".encode(), usedforsecurity=False
+        ).digest()
         return struct.unpack("q", h[:8])[0]
 
 

backend/onyx/chat/chat_state.py

@@ -1,19 +1,8 @@
 import threading
-import time
-from collections.abc import Callable
-from collections.abc import Generator
-from queue import Empty
 
 from onyx.chat.citation_processor import CitationMapping
-from onyx.chat.emitter import Emitter
 from onyx.context.search.models import SearchDoc
-from onyx.server.query_and_chat.placement import Placement
-from onyx.server.query_and_chat.streaming_models import OverallStop
-from onyx.server.query_and_chat.streaming_models import Packet
-from onyx.server.query_and_chat.streaming_models import PacketException
 from onyx.tools.models import ToolCallInfo
-from onyx.utils.threadpool_concurrency import run_in_background
-from onyx.utils.threadpool_concurrency import wait_on_background
 
 # Type alias for search doc deduplication key
 # Simple key: just document_id (str)
@@ -159,114 +148,3 @@ class ChatStateContainer:
         """Thread-safe getter for emitted citations (returns a copy)."""
         with self._lock:
             return self._emitted_citations.copy()
-
-
-def run_chat_loop_with_state_containers(
-    chat_loop_func: Callable[[Emitter, ChatStateContainer], None],
-    completion_callback: Callable[[ChatStateContainer], None],
-    is_connected: Callable[[], bool],
-    emitter: Emitter,
-    state_container: ChatStateContainer,
-) -> Generator[Packet, None]:
-    """
-    Explicit wrapper function that runs a function in a background thread
-    with event streaming capabilities.
-
-    The wrapped function should accept emitter as first arg and use it to emit
-    Packet objects. This wrapper polls every 300ms to check if stop signal is set.
-
-    Args:
-        func: The function to wrap (should accept emitter and state_container as first and second args)
-        completion_callback: Callback function to call when the function completes
-        emitter: Emitter instance for sending packets
-        state_container: ChatStateContainer instance for accumulating state
-        is_connected: Callable that returns False when stop signal is set
-
-    Usage:
-        packets = run_chat_loop_with_state_containers(
-            my_func,
-            completion_callback=completion_callback,
-            emitter=emitter,
-            state_container=state_container,
-            is_connected=check_func,
-        )
-        for packet in packets:
-            # Process packets
-            pass
-    """
-
-    def run_with_exception_capture() -> None:
-        try:
-            chat_loop_func(emitter, state_container)
-        except Exception as e:
-            # If execution fails, emit an exception packet
-            emitter.emit(
-                Packet(
-                    placement=Placement(turn_index=0),
-                    obj=PacketException(type="error", exception=e),
-                )
-            )
-
-    # Run the function in a background thread
-    thread = run_in_background(run_with_exception_capture)
-
-    pkt: Packet | None = None
-    last_turn_index = 0  # Track the highest turn_index seen for stop packet
-    last_cancel_check = time.monotonic()
-    cancel_check_interval = 0.3  # Check for cancellation every 300ms
-    try:
-        while True:
-            # Poll queue with 300ms timeout for natural stop signal checking
-            # the 300ms timeout is to avoid busy-waiting and to allow the stop signal to be checked regularly
-            try:
-                pkt = emitter.bus.get(timeout=0.3)
-            except Empty:
-                if not is_connected():
-                    # Stop signal detected
-                    yield Packet(
-                        placement=Placement(turn_index=last_turn_index + 1),
-                        obj=OverallStop(type="stop", stop_reason="user_cancelled"),
-                    )
-                    break
-                last_cancel_check = time.monotonic()
-                continue
-
-            if pkt is not None:
-                # Track the highest turn_index for the stop packet
-                if pkt.placement and pkt.placement.turn_index > last_turn_index:
-                    last_turn_index = pkt.placement.turn_index
-
-                if isinstance(pkt.obj, OverallStop):
-                    yield pkt
-                    break
-                elif isinstance(pkt.obj, PacketException):
-                    raise pkt.obj.exception
-                else:
-                    yield pkt
-
-                # Check for cancellation periodically even when packets are flowing
-                # This ensures stop signal is checked during active streaming
-                current_time = time.monotonic()
-                if current_time - last_cancel_check >= cancel_check_interval:
-                    if not is_connected():
-                        # Stop signal detected during streaming
-                        yield Packet(
-                            placement=Placement(turn_index=last_turn_index + 1),
-                            obj=OverallStop(type="stop", stop_reason="user_cancelled"),
-                        )
-                        break
-                    last_cancel_check = current_time
-    finally:
-        # Wait for thread to complete on normal exit to propagate exceptions and ensure cleanup.
-        # Skip waiting if user disconnected to exit quickly.
-        if is_connected():
-            wait_on_background(thread)
-        try:
-            completion_callback(state_container)
-        except Exception as e:
-            emitter.emit(
-                Packet(
-                    placement=Placement(turn_index=last_turn_index + 1),
-                    obj=PacketException(type="error", exception=e),
-                )
-            )

backend/onyx/chat/chat_utils.py

@@ -5,6 +5,7 @@ from typing import cast
 from uuid import UUID
 
 from fastapi.datastructures import Headers
+from pydantic import BaseModel
 from sqlalchemy.orm import Session
 
 from onyx.chat.models import ChatHistoryResult
@@ -30,6 +31,8 @@ from onyx.file_processing.extract_file_text import extract_file_text
 from onyx.file_store.file_store import get_default_file_store
 from onyx.file_store.models import ChatFileType
 from onyx.file_store.models import FileDescriptor
+from onyx.file_store.utils import plaintext_file_name_for_id
+from onyx.file_store.utils import store_plaintext
 from onyx.kg.models import KGException
 from onyx.kg.setup.kg_default_entity_definitions import (
     populate_missing_default_entity_types__commit,
@@ -49,6 +52,60 @@ logger = setup_logger()
 IMAGE_GENERATION_TOOL_NAME = "generate_image"
 
 
+class FileContextResult(BaseModel):
+    """Result of building a file's LLM context representation."""
+
+    message: ChatMessageSimple
+    tool_metadata: FileToolMetadata
+
+
+def build_file_context(
+    tool_file_id: str,
+    filename: str,
+    file_type: ChatFileType,
+    content_text: str | None = None,
+    token_count: int = 0,
+    approx_char_count: int | None = None,
+) -> FileContextResult:
+    """Build the LLM context representation for a single file.
+
+    Centralises how files should appear in the LLM prompt
+    — the ID that FileReaderTool accepts (``UserFile.id`` for user files).
+    """
+    if file_type.use_metadata_only():
+        message_text = (
+            f"File: {filename} (id={tool_file_id})\n"
+            "Use the file_reader or python tools to access "
+            "this file's contents."
+        )
+        message = ChatMessageSimple(
+            message=message_text,
+            token_count=max(1, len(message_text) // 4),
+            message_type=MessageType.USER,
+            file_id=tool_file_id,
+        )
+    else:
+        message_text = f"File: {filename}\n{content_text or ''}\nEnd of File"
+        message = ChatMessageSimple(
+            message=message_text,
+            token_count=token_count,
+            message_type=MessageType.USER,
+            file_id=tool_file_id,
+        )
+
+    metadata = FileToolMetadata(
+        file_id=tool_file_id,
+        filename=filename,
+        approx_char_count=(
+            approx_char_count
+            if approx_char_count is not None
+            else len(content_text or "")
+        ),
+    )
+
+    return FileContextResult(message=message, tool_metadata=metadata)
+
+
 def create_chat_session_from_request(
     chat_session_request: ChatSessionCreationRequest,
     user_id: UUID | None,
@@ -289,6 +346,33 @@ def process_kg_commands(
         raise KGException("KG setup done")
 
 
+def _get_or_extract_plaintext(
+    file_id: str,
+    extract_fn: Callable[[], str],
+) -> str:
+    """Load cached plaintext for a file, or extract and store it.
+
+    Tries to read pre-stored plaintext from the file store.  On a miss,
+    calls extract_fn to produce the text, then stores the result so
+    future calls skip the expensive extraction.
+    """
+    file_store = get_default_file_store()
+    plaintext_key = plaintext_file_name_for_id(file_id)
+
+    # Try cached plaintext first.
+    try:
+        plaintext_io = file_store.read_file(plaintext_key, mode="b")
+        return plaintext_io.read().decode("utf-8")
+    except Exception:
+        logger.exception(f"Error when reading file, id={file_id}")
+
+    # Cache miss — extract and store.
+    content_text = extract_fn()
+    if content_text:
+        store_plaintext(file_id, content_text)
+    return content_text
+
+
 @log_function_time(print_only=True)
 def load_chat_file(
     file_descriptor: FileDescriptor, db_session: Session
@@ -303,12 +387,23 @@ def load_chat_file(
     file_type = ChatFileType(file_descriptor["type"])
 
     if file_type.is_text_file():
-        try:
-            content_text = extract_file_text(
+        file_id = file_descriptor["id"]
+
+        def _extract() -> str:
+            return extract_file_text(
                 file=file_io,
                 file_name=file_descriptor.get("name") or "",
                 break_on_unprocessable=False,
             )
+
+        # Use the user_file_id as cache key when available (matches what
+        # the celery indexing worker stores), otherwise fall back to the
+        # file store id (covers code-interpreter-generated files, etc.).
+        user_file_id_str = file_descriptor.get("user_file_id")
+        cache_key = user_file_id_str or file_id
+
+        try:
+            content_text = _get_or_extract_plaintext(cache_key, _extract)
         except Exception as e:
             logger.warning(
                 f"Failed to retrieve content for file {file_descriptor['id']}: {str(e)}"
@@ -498,7 +593,7 @@ def convert_chat_history(
     for idx, chat_message in enumerate(chat_history):
         if chat_message.message_type == MessageType.USER:
             # Process files attached to this message
-            text_files: list[ChatLoadedFile] = []
+            text_files: list[tuple[ChatLoadedFile, FileDescriptor]] = []
             image_files: list[ChatLoadedFile] = []
 
             if chat_message.files:
@@ -509,34 +604,26 @@ def convert_chat_history(
                         if loaded_file.file_type == ChatFileType.IMAGE:
                             image_files.append(loaded_file)
                         else:
-                            # Text files (DOC, PLAIN_TEXT, CSV) are added as separate messages
-                            text_files.append(loaded_file)
+                            # Text files (DOC, PLAIN_TEXT, TABULAR) are added as separate messages
+                            text_files.append((loaded_file, file_descriptor))
 
             # Add text files as separate messages before the user message.
             # Each message is tagged with ``file_id`` so that forgotten files
             # can be detected after context-window truncation.
-            for text_file in text_files:
-                file_text = text_file.content_text or ""
-                filename = text_file.filename
-                message = (
-                    f"File: {filename}\n{file_text}\nEnd of File"
-                    if filename
-                    else file_text
-                )
-                simple_messages.append(
-                    ChatMessageSimple(
-                        message=message,
-                        token_count=text_file.token_count,
-                        message_type=MessageType.USER,
-                        image_files=None,
-                        file_id=text_file.file_id,
-                    )
-                )
-                all_injected_file_metadata[text_file.file_id] = FileToolMetadata(
-                    file_id=text_file.file_id,
-                    filename=filename or "unknown",
-                    approx_char_count=len(file_text),
+            for text_file, fd in text_files:
+                # Use user_file_id as the FileReaderTool accepts that.
+                # Fall back to the file-store path id.
+                tool_id = fd.get("user_file_id") or text_file.file_id
+                filename = text_file.filename or "unknown"
+                ctx = build_file_context(
+                    tool_file_id=tool_id,
+                    filename=filename,
+                    file_type=text_file.file_type,
+                    content_text=text_file.content_text,
+                    token_count=text_file.token_count,
                 )
+                simple_messages.append(ctx.message)
+                all_injected_file_metadata[tool_id] = ctx.tool_metadata
 
             # Sum token counts from image files (excluding project image files)
             image_token_count = (

backend/onyx/chat/emitter.py

@@ -1,19 +1,40 @@
+import threading
 from queue import Queue
 
+from onyx.server.query_and_chat.placement import Placement
 from onyx.server.query_and_chat.streaming_models import Packet
 
 
 class Emitter:
-    """Use this inside tools to emit arbitrary UI progress."""
+    """Routes packets from LLM/tool execution to the ``_run_models`` drain loop.
 
-    def __init__(self, bus: Queue):
-        self.bus = bus
+    Tags every packet with ``model_index`` and places it on ``merged_queue``
+    as a ``(model_idx, packet)`` tuple for ordered consumption downstream.
+
+    Args:
+        merged_queue: Shared queue owned by ``_run_models``.
+        model_idx: Index embedded in packet placements (``0`` for N=1 runs).
+        drain_done: Optional event set by ``_run_models`` when the drain loop
+            exits early (e.g. HTTP disconnect). When set, ``emit`` returns
+            immediately so worker threads can exit fast.
+    """
+
+    def __init__(
+        self,
+        merged_queue: Queue[tuple[int, Packet | Exception | object]],
+        model_idx: int = 0,
+        drain_done: threading.Event | None = None,
+    ) -> None:
+        self._model_idx = model_idx
+        self._merged_queue = merged_queue
+        self._drain_done = drain_done
 
     def emit(self, packet: Packet) -> None:
-        self.bus.put(packet)  # Thread-safe
-
-
-def get_default_emitter() -> Emitter:
-    bus: Queue[Packet] = Queue()
-    emitter = Emitter(bus)
-    return emitter
+        if self._drain_done is not None and self._drain_done.is_set():
+            return
+        base = packet.placement or Placement(turn_index=0)
+        tagged = Packet(
+            placement=base.model_copy(update={"model_index": self._model_idx}),
+            obj=packet.obj,
+        )
+        self._merged_queue.put((self._model_idx, tagged))

backend/onyx/chat/llm_loop.py

@@ -36,9 +36,11 @@ from onyx.db.memory import add_memory
 from onyx.db.memory import update_memory_at_index
 from onyx.db.memory import UserMemoryContext
 from onyx.db.models import Persona
+from onyx.llm.constants import LlmProviderNames
 from onyx.llm.interfaces import LLM
 from onyx.llm.interfaces import LLMUserIdentity
 from onyx.llm.interfaces import ToolChoiceOptions
+from onyx.llm.utils import is_true_openai_model
 from onyx.prompts.chat_prompts import IMAGE_GEN_REMINDER
 from onyx.prompts.chat_prompts import OPEN_URL_REMINDER
 from onyx.server.query_and_chat.placement import Placement
@@ -72,6 +74,70 @@ from shared_configs.contextvars import get_current_tenant_id
 logger = setup_logger()
 
 
+class EmptyLLMResponseError(RuntimeError):
+    """Raised when the streamed LLM response completes without a usable answer."""
+
+    def __init__(
+        self,
+        *,
+        provider: str,
+        model: str,
+        tool_choice: ToolChoiceOptions,
+        client_error_msg: str,
+        error_code: str = "EMPTY_LLM_RESPONSE",
+        is_retryable: bool = True,
+    ) -> None:
+        super().__init__(client_error_msg)
+        self.provider = provider
+        self.model = model
+        self.tool_choice = tool_choice
+        self.client_error_msg = client_error_msg
+        self.error_code = error_code
+        self.is_retryable = is_retryable
+
+
+def _build_empty_llm_response_error(
+    llm: LLM,
+    llm_step_result: LlmStepResult,
+    tool_choice: ToolChoiceOptions,
+) -> EmptyLLMResponseError:
+    provider = llm.config.model_provider
+    model = llm.config.model_name
+
+    # OpenAI quota exhaustion has reached us as a streamed "stop" with zero content.
+    # When the stream is completely empty and there is no reasoning/tool output, surface
+    # the likely account-level cause instead of a generic tool-calling error.
+    if (
+        not llm_step_result.reasoning
+        and provider == LlmProviderNames.OPENAI
+        and is_true_openai_model(provider, model)
+    ):
+        return EmptyLLMResponseError(
+            provider=provider,
+            model=model,
+            tool_choice=tool_choice,
+            client_error_msg=(
+                "The selected OpenAI model returned an empty streamed response "
+                "before producing any tokens. This commonly happens when the API "
+                "key or project has no remaining quota or billing is not enabled. "
+                "Verify quota and billing for this key and try again."
+            ),
+            error_code="BUDGET_EXCEEDED",
+            is_retryable=False,
+        )
+
+    return EmptyLLMResponseError(
+        provider=provider,
+        model=model,
+        tool_choice=tool_choice,
+        client_error_msg=(
+            "The selected model returned no final answer before the stream "
+            "completed. No text or tool calls were received from the upstream "
+            "provider."
+        ),
+    )
+
+
 def _looks_like_xml_tool_call_payload(text: str | None) -> bool:
     """Detect XML-style marshaled tool calls emitted as plain text."""
     if not text:
@@ -613,7 +679,12 @@ def run_llm_loop(
             )
             citation_processor.update_citation_mapping(project_citation_mapping)
 
-        llm_step_result: LlmStepResult | None = None
+        llm_step_result = LlmStepResult(
+            reasoning=None,
+            answer=None,
+            tool_calls=None,
+            raw_answer=None,
+        )
 
         # Pass the total budget to construct_message_history, which will handle token allocation
         available_tokens = llm.config.max_input_tokens
@@ -1084,12 +1155,18 @@ def run_llm_loop(
                 # As long as 1 tool with citeable documents is called at any point, we ask the LLM to try to cite
                 should_cite_documents = True
 
-        if not llm_step_result or not llm_step_result.answer:
+        if not llm_step_result.answer and not llm_step_result.tool_calls:
+            raise _build_empty_llm_response_error(
+                llm=llm,
+                llm_step_result=llm_step_result,
+                tool_choice=tool_choice,
+            )
+
+        if not llm_step_result.answer:
             raise RuntimeError(
-                "The LLM did not return an answer. "
-                "Typically this is an issue with LLMs that do not support tool calling natively, "
-                "or the model serving API is not configured correctly. "
-                "This may also happen with models that are lower quality outputting invalid tool calls."
+                "The LLM did not return a final answer after tool execution. "
+                "Typically this indicates invalid tool-call output, a model/provider mismatch, "
+                "or serving API misconfiguration."
             )
 
         emitter.emit(

backend/onyx/chat/llm_step.py

@@ -1013,6 +1013,10 @@ def run_llm_step_pkt_generator(
     accumulated_reasoning = ""
     accumulated_answer = ""
     accumulated_raw_answer = ""
+    stream_chunk_count = 0
+    actionable_chunk_count = 0
+    empty_chunk_count = 0
+    finish_reasons: set[str] = set()
     xml_tool_call_content_filter = _XmlToolCallContentFilter()
 
     processor_state: Any = None
@@ -1145,6 +1149,7 @@ def run_llm_step_pkt_generator(
             user_identity=user_identity,
             timeout_override=timeout_override,
         ):
+            stream_chunk_count += 1
             if packet.usage:
                 usage = packet.usage
                 span_generation.span_data.usage = {
@@ -1154,16 +1159,21 @@ def run_llm_step_pkt_generator(
                     "cache_creation_input_tokens": usage.cache_creation_input_tokens,
                 }
                 # Note: LLM cost tracking is now handled in multi_llm.py
+            finish_reason = packet.choice.finish_reason
+            if finish_reason:
+                finish_reasons.add(str(finish_reason))
             delta = packet.choice.delta
 
             # Weird behavior from some model providers, just log and ignore for now
             if (
-                delta.content is None
+                not delta.content
                 and delta.reasoning_content is None
-                and delta.tool_calls is None
+                and not delta.tool_calls
             ):
+                empty_chunk_count += 1
                 logger.warning(
-                    f"LLM packet is empty (no contents, reasoning or tool calls). Skipping: {packet}"
+                    "LLM packet is empty (no content, reasoning, or tool calls). "
+                    f"finish_reason={finish_reason}. Skipping: {packet}"
                 )
                 continue
 
@@ -1172,6 +1182,8 @@ def run_llm_step_pkt_generator(
                     time.monotonic() - stream_start_time
                 )
                 first_action_recorded = True
+            if _delta_has_action(delta):
+                actionable_chunk_count += 1
 
             if custom_token_processor:
                 # The custom token processor can modify the deltas for specific custom logic
@@ -1307,6 +1319,15 @@ def run_llm_step_pkt_generator(
         else:
             logger.debug("Tool calls: []")
 
+    if actionable_chunk_count == 0:
+        logger.warning(
+            "LLM stream completed with no actionable deltas. "
+            f"chunks={stream_chunk_count}, empty_chunks={empty_chunk_count}, "
+            f"finish_reasons={sorted(finish_reasons)}, "
+            f"provider={llm.config.model_provider}, model={llm.config.model_name}, "
+            f"tool_choice={tool_choice}, tools_sent={len(tool_definitions)}"
+        )
+
     return (
         LlmStepResult(
             reasoning=accumulated_reasoning if accumulated_reasoning else None,

backend/onyx/chat/models.py

@@ -8,6 +8,7 @@ from onyx.configs.constants import MessageType
 from onyx.context.search.models import SearchDoc
 from onyx.file_store.models import InMemoryChatFile
 from onyx.server.query_and_chat.models import MessageResponseIDInfo
+from onyx.server.query_and_chat.models import MultiModelMessageResponseIDInfo
 from onyx.server.query_and_chat.streaming_models import CitationInfo
 from onyx.server.query_and_chat.streaming_models import GeneratedImage
 from onyx.server.query_and_chat.streaming_models import Packet
@@ -35,7 +36,13 @@ class CreateChatSessionID(BaseModel):
     chat_session_id: UUID
 
 
-AnswerStreamPart = Packet | MessageResponseIDInfo | StreamingError | CreateChatSessionID
+AnswerStreamPart = (
+    Packet
+    | MessageResponseIDInfo
+    | MultiModelMessageResponseIDInfo
+    | StreamingError
+    | CreateChatSessionID
+)
 
 AnswerStream = Iterator[AnswerStreamPart]
 
@@ -177,8 +184,8 @@ class ExtractedContextFiles(BaseModel):
 class SearchParams(BaseModel):
     """Resolved search filter IDs and search-tool usage for a chat turn."""
 
-    search_project_id: int | None
-    search_persona_id: int | None
+    project_id_filter: int | None
+    persona_id_filter: int | None
     search_usage: SearchToolUsage
 
 

backend/onyx/configs/app_configs.py

@@ -44,6 +44,31 @@ SEND_USER_METADATA_TO_LLM_PROVIDER = (
 # User Facing Features Configs
 #####
 BLURB_SIZE = 128  # Number Encoder Tokens included in the chunk blurb
+
+# Hard ceiling for the admin-configurable file upload size (in MB).
+# Self-hosted customers can raise or lower this via the environment variable.
+_raw_max_upload_size_mb = int(os.environ.get("MAX_ALLOWED_UPLOAD_SIZE_MB", "250"))
+if _raw_max_upload_size_mb < 0:
+    logger.warning(
+        "MAX_ALLOWED_UPLOAD_SIZE_MB=%d is negative; falling back to 250",
+        _raw_max_upload_size_mb,
+    )
+    _raw_max_upload_size_mb = 250
+MAX_ALLOWED_UPLOAD_SIZE_MB = _raw_max_upload_size_mb
+
+# Default fallback for the per-user file upload size limit (in MB) when no
+# admin-configured value exists.  Clamped to MAX_ALLOWED_UPLOAD_SIZE_MB at
+# runtime so this never silently exceeds the hard ceiling.
+_raw_default_upload_size_mb = int(
+    os.environ.get("DEFAULT_USER_FILE_MAX_UPLOAD_SIZE_MB", "100")
+)
+if _raw_default_upload_size_mb < 0:
+    logger.warning(
+        "DEFAULT_USER_FILE_MAX_UPLOAD_SIZE_MB=%d is negative; falling back to 100",
+        _raw_default_upload_size_mb,
+    )
+    _raw_default_upload_size_mb = 100
+DEFAULT_USER_FILE_MAX_UPLOAD_SIZE_MB = _raw_default_upload_size_mb
 GENERATIVE_MODEL_ACCESS_CHECK_FREQ = int(
     os.environ.get("GENERATIVE_MODEL_ACCESS_CHECK_FREQ") or 86400
 )  # 1 day
@@ -61,17 +86,6 @@ CACHE_BACKEND = CacheBackendType(
     os.environ.get("CACHE_BACKEND", CacheBackendType.REDIS)
 )
 
-# Maximum token count for a single uploaded file. Files exceeding this are rejected.
-# Defaults to 100k tokens (or 10M when vector DB is disabled).
-_DEFAULT_FILE_TOKEN_LIMIT = 10_000_000 if DISABLE_VECTOR_DB else 100_000
-FILE_TOKEN_COUNT_THRESHOLD = int(
-    os.environ.get("FILE_TOKEN_COUNT_THRESHOLD", str(_DEFAULT_FILE_TOKEN_LIMIT))
-)
-
-# Maximum upload size for a single user file (chat/projects) in MB.
-USER_FILE_MAX_UPLOAD_SIZE_MB = int(os.environ.get("USER_FILE_MAX_UPLOAD_SIZE_MB") or 50)
-USER_FILE_MAX_UPLOAD_SIZE_BYTES = USER_FILE_MAX_UPLOAD_SIZE_MB * 1024 * 1024
-
 # If set to true, will show extra/uncommon connectors in the "Other" category
 SHOW_EXTRA_CONNECTORS = os.environ.get("SHOW_EXTRA_CONNECTORS", "").lower() == "true"
 
@@ -278,14 +292,17 @@ USING_AWS_MANAGED_OPENSEARCH = (
 OPENSEARCH_PROFILING_DISABLED = (
     os.environ.get("OPENSEARCH_PROFILING_DISABLED", "").lower() == "true"
 )
-
+# Whether to disable match highlights for OpenSearch. Defaults to True for now
+# as we investigate query performance.
+OPENSEARCH_MATCH_HIGHLIGHTS_DISABLED = (
+    os.environ.get("OPENSEARCH_MATCH_HIGHLIGHTS_DISABLED", "true").lower() == "true"
+)
 # When enabled, OpenSearch returns detailed score breakdowns for each hit.
 # Useful for debugging and tuning search relevance. Has ~10-30% performance overhead according to documentation.
 # Seems for Hybrid Search in practice, the impact is actually more like 1000x slower.
 OPENSEARCH_EXPLAIN_ENABLED = (
     os.environ.get("OPENSEARCH_EXPLAIN_ENABLED", "").lower() == "true"
 )
-
 # Analyzer used for full-text fields (title, content). Use OpenSearch built-in analyzer
 # names (e.g. "english", "standard", "german"). Affects stemming and tokenization;
 # existing indices need reindexing after a change.
@@ -318,8 +335,20 @@ VERIFY_CREATE_OPENSEARCH_INDEX_ON_INIT_MT = (
 OPENSEARCH_MIGRATION_GET_VESPA_CHUNKS_PAGE_SIZE = int(
     os.environ.get("OPENSEARCH_MIGRATION_GET_VESPA_CHUNKS_PAGE_SIZE") or 500
 )
-OPENSEARCH_OVERRIDE_DEFAULT_NUM_HYBRID_SEARCH_CANDIDATES = int(
-    os.environ.get("OPENSEARCH_DEFAULT_NUM_HYBRID_SEARCH_CANDIDATES") or 0
+# If set, will override the default number of shards and replicas for the index.
+OPENSEARCH_INDEX_NUM_SHARDS: int | None = (
+    int(os.environ["OPENSEARCH_INDEX_NUM_SHARDS"])
+    if os.environ.get("OPENSEARCH_INDEX_NUM_SHARDS", None) is not None
+    else None
+)
+OPENSEARCH_INDEX_NUM_REPLICAS: int | None = (
+    int(os.environ["OPENSEARCH_INDEX_NUM_REPLICAS"])
+    if os.environ.get("OPENSEARCH_INDEX_NUM_REPLICAS", None) is not None
+    else None
+)
+ONYX_SEARCH_UI_USES_OPENSEARCH_KEYWORD_SEARCH = (
+    os.environ.get("ONYX_SEARCH_UI_USES_OPENSEARCH_KEYWORD_SEARCH", "").lower()
+    == "true"
 )
 
 VESPA_HOST = os.environ.get("VESPA_HOST") or "localhost"
@@ -776,6 +805,10 @@ MINI_CHUNK_SIZE = 150
 # This is the number of regular chunks per large chunk
 LARGE_CHUNK_RATIO = 4
 
+# The maximum number of chunks that can be held for 1 document processing batch
+# The purpose of this is to set an upper bound on memory usage
+MAX_CHUNKS_PER_DOC_BATCH = int(os.environ.get("MAX_CHUNKS_PER_DOC_BATCH") or 1000)
+
 # Include the document level metadata in each chunk. If the metadata is too long, then it is thrown out
 # We don't want the metadata to overwhelm the actual contents of the chunk
 SKIP_METADATA_IN_CHUNK = os.environ.get("SKIP_METADATA_IN_CHUNK", "").lower() == "true"
@@ -1046,6 +1079,7 @@ POD_NAMESPACE = os.environ.get("POD_NAMESPACE")
 
 DEV_MODE = os.environ.get("DEV_MODE", "").lower() == "true"
 
+
 INTEGRATION_TESTS_MODE = os.environ.get("INTEGRATION_TESTS_MODE", "").lower() == "true"
 
 #####

backend/onyx/configs/chat_configs.py

@@ -24,11 +24,11 @@ CONTEXT_CHUNKS_BELOW = int(os.environ.get("CONTEXT_CHUNKS_BELOW") or 1)
 LLM_SOCKET_READ_TIMEOUT = int(
     os.environ.get("LLM_SOCKET_READ_TIMEOUT") or "60"
 )  # 60 seconds
-# Weighting factor between Vector and Keyword Search, 1 for completely vector search
+# Weighting factor between vector and keyword Search; 1 for completely vector
+# search, 0 for keyword. Enforces a valid range of [0, 1]. A supplied value from
+# the env outside of this range will be clipped to the respective end of the
+# range. Defaults to 0.5.
 HYBRID_ALPHA = max(0, min(1, float(os.environ.get("HYBRID_ALPHA") or 0.5)))
-HYBRID_ALPHA_KEYWORD = max(
-    0, min(1, float(os.environ.get("HYBRID_ALPHA_KEYWORD") or 0.4))
-)
 # Weighting factor between Title and Content of documents during search, 1 for completely
 # Title based. Default heavily favors Content because Title is also included at the top of
 # Content. This is to avoid cases where the Content is very relevant but it may not be clear

backend/onyx/configs/constants.py

@@ -177,6 +177,14 @@ USER_FILE_PROJECT_SYNC_MAX_QUEUE_DEPTH = 500
 
 CELERY_USER_FILE_PROJECT_SYNC_LOCK_TIMEOUT = 5 * 60  # 5 minutes (in seconds)
 
+# How long a queued user-file-delete task is valid before workers discard it.
+# Mirrors the processing task expiry to prevent indefinite queue growth when
+# files are stuck in DELETING status and the beat keeps re-enqueuing them.
+CELERY_USER_FILE_DELETE_TASK_EXPIRES = 60  # 1 minute (in seconds)
+
+# Max queue depth before the delete beat stops enqueuing more delete tasks.
+USER_FILE_DELETE_MAX_QUEUE_DEPTH = 500
+
 CELERY_SANDBOX_FILE_SYNC_LOCK_TIMEOUT = 5 * 60  # 5 minutes (in seconds)
 
 DANSWER_REDIS_FUNCTION_LOCK_PREFIX = "da_function_lock:"
@@ -204,6 +212,7 @@ class DocumentSource(str, Enum):
     PRODUCTBOARD = "productboard"
     FILE = "file"
     CODA = "coda"
+    CANVAS = "canvas"
     NOTION = "notion"
     ZULIP = "zulip"
     LINEAR = "linear"
@@ -469,6 +478,9 @@ class OnyxRedisLocks:
     USER_FILE_PROJECT_SYNC_QUEUED_PREFIX = "da_lock:user_file_project_sync_queued"
     USER_FILE_DELETE_BEAT_LOCK = "da_lock:check_user_file_delete_beat"
     USER_FILE_DELETE_LOCK_PREFIX = "da_lock:user_file_delete"
+    # Short-lived key set when a delete task is enqueued; cleared when the worker picks it up.
+    # Prevents the beat from re-enqueuing the same file while a delete task is already queued.
+    USER_FILE_DELETE_QUEUED_PREFIX = "da_lock:user_file_delete_queued"
 
     # Release notes
     RELEASE_NOTES_FETCH_LOCK = "da_lock:release_notes_fetch"
@@ -597,6 +609,9 @@ class OnyxCeleryTask:
     EXPORT_QUERY_HISTORY_TASK = "export_query_history_task"
     EXPORT_QUERY_HISTORY_CLEANUP_TASK = "export_query_history_cleanup_task"
 
+    # Hook execution log retention
+    HOOK_EXECUTION_LOG_CLEANUP_TASK = "hook_execution_log_cleanup_task"
+
     # Sandbox cleanup
     CLEANUP_IDLE_SANDBOXES = "cleanup_idle_sandboxes"
     CLEANUP_OLD_SNAPSHOTS = "cleanup_old_snapshots"
@@ -658,6 +673,7 @@ DocumentSourceDescription: dict[DocumentSource, str] = {
     DocumentSource.SLAB: "slab data",
     DocumentSource.PRODUCTBOARD: "productboard data (boards, etc.)",
     DocumentSource.FILE: "files",
+    DocumentSource.CANVAS: "canvas lms - courses, pages, assignments, and announcements",
     DocumentSource.CODA: "coda - team workspace with docs, tables, and pages",
     DocumentSource.NOTION: "notion data - a workspace that combines note-taking, \
 project management, and collaboration tools into a single, customizable platform",