[jira] [Commented] (FLINK-15356) Add applicationId to existing flink metrics running on yarn

["Chesnay Schepler (Jira)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/FLINK-15356?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17047428#comment-17047428 ] 

Chesnay Schepler commented on FLINK-15356:
------------------------------------------

This needs a larger discussion on identifying Flink clusters; in it's current form the PR introduces a yarn-specific feature with no way of using it if you're not on yarn, and even introduces yarn-specific code into the metric system which just shouldn't happen.

Doing this type of change in isolation doesn't make sense; if we want clusters to be indentifiable on some way then it should be used exposed via metrics, logging and WebUI. Maybe even the client.

> Add applicationId to existing flink metrics running on yarn
> -----------------------------------------------------------
>
>                 Key: FLINK-15356
>                 URL: https://issues.apache.org/jira/browse/FLINK-15356
>             Project: Flink
>          Issue Type: Improvement
>          Components: Runtime / Metrics
>            Reporter: Forward Xu
>            Priority: Major
>              Labels: pull-request-available
>          Time Spent: 10m
>  Remaining Estimate: 0h
>
> When sending metrics to Prometheus, these systems have only the Flink job ID, and the Flink job ID is UUID, which cannot be associated with the application job on the yarn. Therefore, we need to increase the applicationId when running on yarn. This helps us to accurately find the corresponding job on yarn when the metric is abnormal.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)