zmodz1010 commented on code in PR #76682:
URL: https://github.com/apache/airflow/pull/76682#discussion_r4527386670
##########
airflow-core/src/airflow/models/dagrun.py:
##########
@@ -1638,6 +1638,28 @@ def _build_finished_tis(dag: SerializedDAG, rows) ->
list[FinishedTI]:
unfinished_tis = list(_filter_tis_and_exclude_removed(dag,
unfinished_tis))
finished_tis: list[SchedulingTI] = list(_build_finished_tis(dag,
finished_rows))
+ retry_timeout_expired = False
+ now = timezone.utcnow()
+ for ti in unfinished_tis:
+ if ti.state not in {
+ TaskInstanceState.UP_FOR_RETRY,
+ TaskInstanceState.SCHEDULED,
+ TaskInstanceState.QUEUED,
+ }:
+ continue
+ if ti.retry_timeout_deadline is None:
+ if ti.task.retry_timeout is None:
+ continue
+ failed_at = ti.end_date
+ if failed_at is None and ti.try_number > 1:
+ failed_at =
ti._get_previous_attempt_end_date(session=session)
+ if failed_at is None:
+ continue
+ ti.retry_timeout_deadline = failed_at + ti.task.retry_timeout
+ if ti.retry_timeout_deadline <= now:
+ ti.set_state(TaskInstanceState.FAILED, session=session)
Review Comment:
Good point. Retry-window expiry previously marked the task `FAILED` without
entering the usual failure-notification path. I’ve added a regression test for
both `on_failure_callback` and `email_on_failure`, and the scheduler now
dispatches those notifications when it first marks a waiting retry as failed.
The fix is pushed in `c133922725`.
Ruff formatting and lint pass. I couldn’t run the focused test locally:
Airflow import fails in this Windows checkout because
`airflow._shared.configuration` is unavailable, so CI will need to validate the
test.
--
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]
