[jira] [Commented] (FLINK-8506) fullRestarts Gauge not incremented when jobmanager got killed
[ https://issues.apache.org/jira/browse/FLINK-8506?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16339560#comment-16339560 ] Steven Zhen Wu commented on FLINK-8506: --- Till, thanks for the explanation. Looks like we should clarify the doc, which says "since job submitted". [https://ci.apache.org/projects/flink/flink-docs-master/monitoring/metrics.html] fullRestarts The total number of full restarts since this job was submitted (in milliseconds). Gauge So it seems that we don't any metric to capture jobmanager failover. > fullRestarts Gauge not incremented when jobmanager got killed > - > > Key: FLINK-8506 > URL: https://issues.apache.org/jira/browse/FLINK-8506 > Project: Flink > Issue Type: Bug >Reporter: Steven Zhen Wu >Priority: Major > > [~till.rohrmann] When jobmanager node got killed, it will cause job restart. > But in this case, we didn't see _fullRestarts_ guage got incremented. is this > expected or a bug? -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (FLINK-8506) fullRestarts Gauge not incremented when jobmanager got killed
[
https://issues.apache.org/jira/browse/FLINK-8506?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16339177#comment-16339177
]
Till Rohrmann commented on FLINK-8506:
--
In case of a JobManager failover the {{fullRestarts}} gauge will be reset since
the number of {{fullRestarts}} is not persisted. This applies to more or less
all metrics in Flink.
> fullRestarts Gauge not incremented when jobmanager got killed
> -
>
> Key: FLINK-8506
> URL: https://issues.apache.org/jira/browse/FLINK-8506
> Project: Flink
> Issue Type: Bug
>Reporter: Steven Zhen Wu
>Priority: Major
>
> [~till.rohrmann] When jobmanager node got killed, it will cause job restart.
> But in this case, we didn't see _fullRestarts_ guage got incremented. is this
> expected or a bug?
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
