[
https://issues.apache.org/jira/browse/YARN-5995?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15753958#comment-15753958
]
Sunil G commented on YARN-5995:
-------------------------------
[~piaoyu zhang]
Thanks for the patch.
Generally I think that its better to discuss what all metrics are we going to
publish from StateStore and the type of the metric etc. I think it will help to
streamline the most needed metrics to be exposed as of priority.
Currently you have exposed {{Duration}} for all type of apis from StateStore
interface. I think this is too much of an information and we need to see
whether we need all such metrics and what we can infer out of same. I think
{{Duration}} overall for all api's may be more meaningful.So we can see how
fast the underlying storage is finishing each api writes and gives a metric
like {{no: of write ops per secs}} etc. Because I am not may not be much
worried about each api's storage duration.
I think we also need data size distribution. We may need to see how much data
is stored over time and its latency. As mentioned in YARN-3936, I also think
that *write latency* and *data size distribution* are major to start with.
> Add RMStateStore metrics to monitor all RMStateStoreEventTypeTransition
> performance
> -----------------------------------------------------------------------------------
>
> Key: YARN-5995
> URL: https://issues.apache.org/jira/browse/YARN-5995
> Project: Hadoop YARN
> Issue Type: Improvement
> Components: metrics, resourcemanager
> Affects Versions: 2.7.1
> Environment: CentOS7.2 Hadoop-2.7.1
> Reporter: zhangyubiao
> Labels: patch
> Attachments: YARN-5995.0001.patch, YARN-5995.0002.patch,
> YARN-5995.patch
>
>
> Add RMStateStore metrics to monitor all RMStateStoreEventTypeTransition
> performance
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
