You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@flink.apache.org by "Chesnay Schepler (JIRA)" <ji...@apache.org> on 2017/01/20 05:41:27 UTC

[jira] [Updated] (FLINK-5150) WebUI metric-related resource leak

     [ https://issues.apache.org/jira/browse/FLINK-5150?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Chesnay Schepler updated FLINK-5150:
------------------------------------
    Affects Version/s:     (was: 1.1.3)

> WebUI metric-related resource leak
> ----------------------------------
>
>                 Key: FLINK-5150
>                 URL: https://issues.apache.org/jira/browse/FLINK-5150
>             Project: Flink
>          Issue Type: Bug
>          Components: Webfrontend
>    Affects Versions: 1.2.0, 1.3.0
>            Reporter: Chesnay Schepler
>            Assignee: Chesnay Schepler
>            Priority: Critical
>             Fix For: 1.2.0
>
>
> The WebUI maintains a list of selected metrics for all jobs and vertices. When a metric is selected in the metric screen it is added to this list, and removed if it is unselected.
> The contents of this list are stored in the browser's localStorage. This allows a user to setup a metric screen, move to another page, and return to the original screen completely intact.
> However, if the metrics are never *unselected* by the user they will remain in this list. They will also still be in this list if the WebUI can't even display the corresponding job page anymore, if for example the history size limit was exceeded. They will even survive a browser restart, since they are not stored in a session-based storage.
> Furthermore, the WebUI still tries to update these metricsd, adding additional overhead to the WebBackend and potentially network.
> In other words, if you _ever_ checked out metrics tab for some job, chances are that the next time you start the WebInterface it will still try to update the metrics for it.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)