timestamp format fix

backend/onyx/connectors/google_drive/connector.py

@@ -1097,7 +1097,9 @@ class GoogleDriveConnector(SlimConnector, CheckpointConnector[GoogleDriveCheckpo
             drive_service.files().list(pageSize=1, fields="files(id)").execute()
 
             if isinstance(self._creds, ServiceAccountCredentials):
-                retry_builder()(get_root_folder_id)(drive_service)
+                # default is ~17mins of retries, don't do that here since this is called from
+                # the UI
+                retry_builder(tries=3, delay=0.1)(get_root_folder_id)(drive_service)
 
         except HttpError as e:
             status_code = e.resp.status if e.resp else None

backend/onyx/connectors/google_drive/file_retrieval.py

@@ -1,6 +1,7 @@
 from collections.abc import Callable
 from collections.abc import Iterator
 from datetime import datetime
+from datetime import timezone
 
 from googleapiclient.discovery import Resource  # type: ignore
 
@@ -36,12 +37,12 @@ def _generate_time_range_filter(
 ) -> str:
     time_range_filter = ""
     if start is not None:
-        time_start = datetime.utcfromtimestamp(start).isoformat() + "Z"
+        time_start = datetime.fromtimestamp(start, tz=timezone.utc).isoformat()
         time_range_filter += (
             f" and {GoogleFields.MODIFIED_TIME.value} >= '{time_start}'"
         )
     if end is not None:
-        time_stop = datetime.utcfromtimestamp(end).isoformat() + "Z"
+        time_stop = datetime.fromtimestamp(end, tz=timezone.utc).isoformat()
         time_range_filter += f" and {GoogleFields.MODIFIED_TIME.value} <= '{time_stop}'"
     return time_range_filter
 

backend/onyx/connectors/google_utils/google_utils.py

@@ -17,9 +17,12 @@ logger = setup_logger()
 
 
 # Google Drive APIs are quite flakey and may 500 for an
-# extended period of time. Trying to combat here by adding a very
-# long retry period (~20 minutes of trying every minute)
-add_retries = retry_builder(tries=50, max_delay=30)
+# extended period of time. This is now addressed by checkpointing.
+#
+# NOTE: We previously tried to combat this here by adding a very
+# long retry period (~20 minutes of trying, one request a minute.)
+# This is no longer necessary due to checkpointing.
+add_retries = retry_builder(tries=5, max_delay=10)
 
 NEXT_PAGE_TOKEN_KEY = "nextPageToken"
 PAGE_TOKEN_KEY = "pageToken"
@@ -37,14 +40,14 @@ class GoogleFields(str, Enum):
 
 
 def _execute_with_retry(request: Any) -> Any:
-    max_attempts = 10
+    max_attempts = 6
     attempt = 1
 
     while attempt < max_attempts:
         # Note for reasons unknown, the Google API will sometimes return a 429
         # and even after waiting the retry period, it will return another 429.
         # It could be due to a few possibilities:
-        # 1. Other things are also requesting from the Gmail API with the same key
+        # 1. Other things are also requesting from the Drive/Gmail API with the same key
         # 2. It's a rolling rate limit so the moment we get some amount of requests cleared, we hit it again very quickly
         # 3. The retry-after has a maximum and we've already hit the limit for the day
         # or it's something else...