[
https://issues.apache.org/jira/browse/YUNIKORN-627?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17334476#comment-17334476
]
Weiwei Yang commented on YUNIKORN-627:
--------------------------------------
I see how this is happening. When the backend collects the data from the
metrics system, there are two different metrics to track num of containers.
# Allocated Containers
# Released Containers
the ACTUAL allocated containers = Allocated - Released. A simple way to fix
this is here:
https://github.com/apache/incubator-yunikorn-core/blob/c220a295d7bc742d898dde96ddea7026dc457bb5/pkg/metrics/metrics_collector.go#L65-L70.
Just need to get released containers, and do a sub before storing it:
u.metricsHistory.Store(totalAppsRunning, totalContainersRunning)
> The container history graph does not go down when pods are gone
> ---------------------------------------------------------------
>
> Key: YUNIKORN-627
> URL: https://issues.apache.org/jira/browse/YUNIKORN-627
> Project: Apache YuniKorn
> Issue Type: Sub-task
> Components: webapp
> Reporter: Chaoran Yu
> Assignee: Wen-Chien,Juan
> Priority: Major
> Attachments: Screen Shot 2021-04-06 at 11.29.26 PM.png
>
>
> In the dashboard view, one of the graphs is the container history. Its value
> keeps increasing even when pods in the cluster come and go. The number in the
> graph should go down when pods are deleted or terminated.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
