[
https://issues.apache.org/jira/browse/FLINK-22164?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17320849#comment-17320849
]
Chesnay Schepler commented on FLINK-22164:
------------------------------------------
[~trohrmann] This is unfortunately not possible; the job id/name are not
available in the JM scope.
https://ci.apache.org/projects/flink/flink-docs-stable/ops/metrics.html#system-scope
The reason is that the JM metrics are actually a mixture of Dispatcher and
JobMaster metrics, and the job id/name could only be known for the latter.
We never adjusted the scopes after FLIP-6.
> Add jobId and JobName variable to JobManager metrics in per-job deployment
> mode
> -------------------------------------------------------------------------------
>
> Key: FLINK-22164
> URL: https://issues.apache.org/jira/browse/FLINK-22164
> Project: Flink
> Issue Type: Improvement
> Components: Runtime / Metrics
> Reporter: Lu Niu
> Priority: Minor
>
> We expose all flink metrics to external system for monitoring and alerting.
> However, JobManager metrics only have one variable <hostname>, which is not
> enough to target to one job when job is deployed to YARN. If flink job runs
> in per-job mode, which ensure one job per cluster, we can add jobId and
> JobName to JobMananger metrics.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
