[jira] [Commented] (FLINK-7069) Catch exceptions for each reporter separately

["ASF GitHub Bot (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/FLINK-7069?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16072572#comment-16072572 ] 

ASF GitHub Bot commented on FLINK-7069:
---------------------------------------

GitHub user zentol opened a pull request:

    https://github.com/apache/flink/pull/4248

    [FLINK-7069] [metrics] Granular exception catching in MetricRegistry

    This PR makes the exception catching in the MetricRegistry more granular when adding or removing metrics.
    
    A reporter that throws an exception prevented other reporters, the MetricQueryService and the ViewUpdater from being notified about added or removed metrics.


You can merge this pull request into a Git repository by running:

    $ git pull https://github.com/zentol/flink 7069

Alternatively you can review and apply these changes as the patch at:

    https://github.com/apache/flink/pull/4248.patch

To close this pull request, make a commit to your master/trunk branch
with (at least) the following in the commit message:

    This closes #4248
    
----
commit 580bdaffdabfedd859e30ee7f37cf3f7983074f9
Author: zentol <ch...@apache.org>
Date:   2017-07-03T14:49:01Z

    [FLINK-7069] [metrics] Granular exception catching in MetricRegistry

----


> Catch exceptions for each reporter separately
> ---------------------------------------------
>
>                 Key: FLINK-7069
>                 URL: https://issues.apache.org/jira/browse/FLINK-7069
>             Project: Flink
>          Issue Type: Improvement
>          Components: Metrics
>    Affects Versions: 1.3.1, 1.4.0
>            Reporter: Chesnay Schepler
>            Assignee: Chesnay Schepler
>             Fix For: 1.4.0
>
>
> The metric system can be effectively disabled by a reporter that throws exceptions whenever it is notified of adding metrics.
> The reason is that the catching of exceptions isn't granular enough, as this peace of psude code shows:
> {code}
> addMetric(metric):
> 	try
> 		for reporter in reporters:
> 			reporter.addMetric(metric)
> 		metricQueryService.addMetric(metric)
> 	catch (e)
> 		logError(e)	
> {code}
> If a reporter throws an exception we never even attempt the other reporters, not notify the MQS which disabled metrics in the WebUI.



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)