[
https://issues.apache.org/jira/browse/FLINK-6477?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16002633#comment-16002633
]
Chesnay Schepler commented on FLINK-6477:
-----------------------------------------
This is an artifact as to how the web-frontend communicates with the backend
for requesting metrics, and how these metrics retrieved from the TaskManagers.
The Frontend requests the metrics from the backend. The backend kicks off an
asynchronous update to the stored metrics, but returns the current state for
the metrics as is, so as to not block the netty thread.
Typically this is only visible on the very first access to any page that relies
on metrics.
How do you intend to fix this? You would have to send an answer to the
web-frontend ansynchronously and I'm not sure if hat is currently possible.
> The first time to click Taskmanager cannot get the actual data
> --------------------------------------------------------------
>
> Key: FLINK-6477
> URL: https://issues.apache.org/jira/browse/FLINK-6477
> Project: Flink
> Issue Type: Bug
> Components: Webfrontend
> Affects Versions: 1.2.0
> Reporter: zhihao chen
> Assignee: zhihao chen
> Attachments: errDisplay.jpg
>
>
> Flink web page first click Taskmanager to get less than the actual data, when
> the parameter “jobmanager.web.refresh-interval” is set to a larger value, eg:
> 1800000, if you do not manually refresh the page you need to wait time after
> the timeout normal display
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
