amoghrajesh commented on code in PR #68213:
URL: https://github.com/apache/airflow/pull/68213#discussion_r3372959807
##########
task-sdk/src/airflow/sdk/bases/resumablejobmixin.py:
##########
@@ -101,29 +107,56 @@ def execute_resumable(self, context: Context) -> Any:
Closing this window would require atomic "submit + persist", which is
not possible across
an external system boundary.
"""
- task_store = context.get("task_store")
+ operator_tag = {"operator": type(self).__name__}
+
+ with tracer.start_as_current_span("resumable_job.resume_decision") as
span:
Review Comment:
The task runner already has a `flush_spans()` context manager wrapping
`main()` that calls `provider.force_flush()` on clean exit, so normal exits are
fine. The missing parent is from a hard crash (SIGKILL/OOM) where
`flush_spans()` never runs and nothing can be done there, spans in the buffer
are gone.
The one gap is SIGTERM: the `_on_term` handler calls `on_kill()` but does
not flush spans before the process exits. Adding a `force_flush()` call there
would recover that case. That feels like a separate task runner improvement
though and is out of scope here. Happy to open a follow-up if you agree.
--
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]
