[
https://issues.apache.org/jira/browse/FLINK-32199?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17729222#comment-17729222
]
Lijie Wang commented on FLINK-32199:
------------------------------------
Fix via:
master(1.18): f3ab9626bf18cad993f7cecba23a7bce6e14407b
release-1.17: 07959a4141ba599194e1e0a8b6da163793d53d0d
release-1.16: e90b2e01fd42330e1115a96e2045d0b3189ba321
> MetricStore does not remove metrics of nonexistent parallelism in
> TaskMetricStore when scale down job parallelism
> -----------------------------------------------------------------------------------------------------------------
>
> Key: FLINK-32199
> URL: https://issues.apache.org/jira/browse/FLINK-32199
> Project: Flink
> Issue Type: Bug
> Components: Runtime / Metrics, Runtime / REST
> Affects Versions: 1.17.0
> Reporter: Junrui Li
> Assignee: Junrui Li
> Priority: Major
> Labels: pull-request-available
> Fix For: 1.18.0, 1.16.3, 1.17.2
>
>
> After FLINK-29615, FLINK will update the subtask metrics store when scaling
> down parallelism. However, task metrics are added in the form of
> "subtaskIndex + metric.name" or "subtaskIndex + operatorName + metric.name".
> Users will be able to find many redundant metrics through
> JobVertexMetricsHandler, which will be very troublesome for users.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
