You are viewing a plain text version of this content. The canonical link for it is here.

Posted to issues@flink.apache.org by "Gyula Fora (Jira)" <ji...@apache.org> on 2024/01/23 12:11:00 UTC

[jira] [Commented] (FLINK-34213) Consider using accumulated busy time instead of busyMsPerSecond

    [ https://issues.apache.org/jira/browse/FLINK-34213?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17809922#comment-17809922 ] 

Gyula Fora commented on FLINK-34213:
------------------------------------

The problem here is that we are using aggregated metrics only currently. To track diff you would need to query individual vertex metrics and track them overhead which would be a huge overhead / cost for jobs with many vertices

> Consider using accumulated busy time instead of busyMsPerSecond
> ---------------------------------------------------------------
>
>                 Key: FLINK-34213
>                 URL: https://issues.apache.org/jira/browse/FLINK-34213
>             Project: Flink
>          Issue Type: Improvement
>          Components: Autoscaler, Kubernetes Operator
>            Reporter: Maximilian Michels
>            Priority: Minor
>
> We might achieve much better accuracy if we used the accumulated busy time metrics from Flink, instead of the momentarily collected ones.
> We would use the diff between the last accumulated and the current accumulated busy time.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)