[
https://issues.apache.org/jira/browse/BEAM-8422?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16953680#comment-16953680
]
Lukasz Gajowy edited comment on BEAM-8422 at 10/17/19 12:09 PM:
----------------------------------------------------------------
That's a fair point - I agree with you. Looking more generally, I think that
ANY job status from a phrase-triggered job SHOULD NOT be notified on builds@.
The way I understand builds@ is that it's a place for builds from master only.
On the contrary, PR triggers status is only for dev+reviewers information so
it's not a thing that should be shared to whole mailing list imo.
was (Author: łukaszg):
That's a fair point - I agree with you. Looking more generally, I think that
ANY job status from a phrase-triggered job SHOULD NOT be notified on builds@.
The way I understand builds@ is that it's a place for builds from master only.
> Send emails to builds@ when job ends with state "ABORTED"
> ---------------------------------------------------------
>
> Key: BEAM-8422
> URL: https://issues.apache.org/jira/browse/BEAM-8422
> Project: Beam
> Issue Type: Improvement
> Components: testing
> Reporter: Lukasz Gajowy
> Priority: Minor
> Time Spent: 20m
> Remaining Estimate: 0h
>
> I noticed that timeouting jobs end up in an ABORTED state (black dot on the
> Jenkins dashboard). No email is sent to the builds@ list when this happens.
> It reduces the visibility of a problem - if one relies on builds@ they won't
> see that Jenkins job takes too much time.
>
> At the time of writing, 5 jobs are affected:
> beam_PostCommit_Java11_ValidatesRunner_Dataflow_PR
> beam_PostCommit_Java11_ValidatesRunner_PortabilityApi_Dataflow
> beam_PostCommit_Java_PVR_Spark_Batch
> beam_PostCommit_Python37_PR
> beam_sonarqube_report
>
> I propose changing this behavior and send emails to builds@ when job ends
> with ABORTED state.
> The drawback of this solution is that every time someone aborts the job
> manually the email will be sent too - there's no way to distinguish those
> two situations. However, IMO we should not allow timeouts to be unnoticed and
> manual job aborting does not happen very often (even committers cannot do
> that now in Jenkins).
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
