[
https://issues.apache.org/jira/browse/YUNIKORN-3112?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Wilfred Spiegelenburg updated YUNIKORN-3112:
--------------------------------------------
Target Version: 1.8.0
> Enhance rate limiting of unschedulable pod events
> -------------------------------------------------
>
> Key: YUNIKORN-3112
> URL: https://issues.apache.org/jira/browse/YUNIKORN-3112
> Project: Apache YuniKorn
> Issue Type: Bug
> Components: core - scheduler
> Reporter: Peter Bacsko
> Assignee: Peter Bacsko
> Priority: Major
>
> Sending out events in every 15 seconds for unschedulable pods can result in
> performance degradation of etcd if the number of such pods are large.
> Possible approaches:
> * More aggressive rate limiting, eg. 5 minutes instead of 15 seconds
> * Do not send an update at all if the event message is not changing
> * Do not send large messages - just send the failed plugin instead of a long
> diagnostic message (eg. "TaintToleration (4x)" instead of a longer "node(s)
> had untolerated taint \{xyz: abc} (4x); failed plugin: 'TaintToleration'")
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
