dabla opened a new pull request, #77202:
URL: https://github.com/apache/airflow/pull/77202

   `KiotaRequestAdapterHook.get_async_conn` reads the adapter cache, then 
awaits the connection, then builds an adapter and writes it to the cache. 
Coroutines that use the same connection for the first time at once (several 
triggers starting together, or async tasks sharing an event loop) all read an 
empty cache while the first one is still awaiting the connection, so each 
builds its own request adapter. Only the last one stays in the cache.
   
   The other adapters still serve the request that built them, but `close()` 
can no longer reach them: it closes the adapter it pops from the cache. Their 
credential sessions are therefore never closed 
(`CachedAsyncTokenCredential.close()` is a no-op by design), and asyncio logs 
this at error level when they are garbage collected, in the middle of the task:
   
   ```
   Unclosed client session
   client_session: <aiohttp.client.ClientSession object at 0x...>
   Unclosed connector
   ```
   
   Each extra adapter also requests its own access token and opens its own 
connection to the API. In task logs this shows as N `Creating Microsoft Graph 
SDK client` blocks followed by N-1 `Unclosed client session` errors.
   
   The cache is now read again after the awaits. Nothing awaits between that 
lookup and the write, so one adapter is built per connection and every caller 
gets the cached one.
   
   The new test gathers five `get_async_conn` calls on an empty cache. Without 
the fix it fails with `Expected '_build_request_adapter' to have been called 
once. Called 5 times.`
   
   Checks run locally: `pytest 
providers/microsoft/azure/tests/unit/microsoft/azure/hooks/test_msgraph.py` 
(121 passed, 1 xfailed) and `prek run --files` on the two changed files.
   
   ---
   
   ##### Was generative AI tooling used to co-author this PR?
   
   - [X] Yes (please specify the tool below)
   
   Generated-by: Claude Code (Fable 5.2) following [the 
guidelines](https://github.com/apache/airflow/blob/main/contributing-docs/05_pull_requests.rst#gen-ai-assisted-contributions)
   
   ---
   
   * Read the **[Pull Request 
Guidelines](https://github.com/apache/airflow/blob/main/contributing-docs/05_pull_requests.rst#pull-request-guidelines)**
 for more information. Note: commit author/co-author name and email in commits 
become permanently public when merged.
   * For fundamental code changes, an Airflow Improvement Proposal 
([AIP](https://cwiki.apache.org/confluence/display/AIRFLOW/Airflow+Improvement+Proposals))
 is needed.
   * When adding dependency, check compliance with the [ASF 3rd Party License 
Policy](https://www.apache.org/legal/resolved.html#category-x).
   * For significant user-facing changes create newsfragment: 
`{pr_number}.significant.rst`, in 
[airflow-core/newsfragments](https://github.com/apache/airflow/tree/main/airflow-core/newsfragments).
 You can add this file in a follow-up commit after the PR is created so you 
know the PR number.
   


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to