-
Notifications
You must be signed in to change notification settings - Fork 14.2k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
execution_timeout
seems to be ignored in EcsRunTaskOperator
when using deferable
#32580
Comments
#31881 was the PR that introduced this. @vandonr-amz are you able to determine whether this is indeed a bug? |
Hmm that's interesting because I'm also confused by what's the expected behavior in this report. Setting a large I actually see 3 bugs being reported:
|
… of failing reported by user in issue apache#32580 the issue is about something else, but the user mentionned this as a "bonus bug"
After testing it, I can confirm that there was a bug where success was returned after max_attempts had been reached however, setting a short and setting a long execution timeout works well too, but it doesn't "override" the max attempts. Maybe another way to see this is that the defaults that are set for the ecs operator are too short, and we should increase those values so that less people would run into timeouts "out of the box". |
This issue has been automatically marked as stale because it has been open for 30 days with no response from the author. It will be closed in next 7 days if no further activity occurs from the issue author. |
It seems this #32589 pr has fixed the fact that the task doesn't return success anymore, however - it doesn't seem to have solved the "killing of the task part". |
ah ok, I missed the |
using the search, it seems like it was reported a looong time ago in #19929 (though at the same time as an other issue too) |
Here is the generic issue for the fact that There are also some workarounds described in that issue for provider/operator developers before its fully fixed. |
Apache Airflow version
2.6.3
What happened
It seems that with the new
aws
provider package, when using thedeferable
keyword in theEcsRunTaskOperator
- theexecution_timeout
is ignored and the task is killed from another timeout, the trigger timeout seems to betimeout=timedelta(seconds=self.waiter_max_attempts * self.waiter_delay + 60)
.Also, it seems when the
trigger
fires that timeout - it seems the task return "success" even though it hasn't finished.It seems this doesn't kill the task either.
What you think should happen instead
The
execution_timeout
should be used in the trigger timeout, or at least a warning if that timeout is overriden or is smallerHow to reproduce
Run an
EcsRunTaskOperator
task with deferable mode, put a largeexecution_timeout
and a small number ofwaiter_retries
. The task should terminates based on thetrigger
timing out before theexecution_timeout
is up.Operating System
linux ubuntu
Versions of Apache Airflow Providers
No response
Deployment
Other Docker-based deployment
Deployment details
No response
Anything else
No response
Are you willing to submit PR?
Code of Conduct
The text was updated successfully, but these errors were encountered: