[ 
https://issues.apache.org/jira/browse/YARN-8692?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16589730#comment-16589730
 ] 

Tao Yang commented on YARN-8692:
--------------------------------

Thanks [~cheersyang] for the feedback. 
Attached v1 patch, please help to review in your free time.

> Support node utilization metrics for SLS
> ----------------------------------------
>
>                 Key: YARN-8692
>                 URL: https://issues.apache.org/jira/browse/YARN-8692
>             Project: Hadoop YARN
>          Issue Type: Improvement
>          Components: scheduler-load-simulator
>    Affects Versions: 3.2.0
>            Reporter: Tao Yang
>            Assignee: Tao Yang
>            Priority: Major
>         Attachments: YARN-8692.001.patch, image-2018-08-21-18-04-22-749.png
>
>
> The distribution of node utilization is an important healthy factor for the 
> YARN cluster, related metrics in SLS can be used to evaluate the scheduling 
> effects and optimize related configurations. 
> To implement this improvement, we need to do things as below:
> (1) Add input configurations (contain avg and stddev for cpu/memory 
> utilization ratio) and generate utilization samples for tasks, not include AM 
> container cause I think it's negligible.
> (2) Simulate containers and node utilization within node status. 
> (3) calculate and generate the distribution metrics and use standard 
> deviation metric (stddev for short) to evaluate the effects(smaller is 
> better).  
> (4) show these metrics on SLS simulator page like this:
> !image-2018-08-21-18-04-22-749.png!
> For Node memory/CPU utilization distribution graphs, Y-axis is nodes number, 
> and P0 represents 0%~9% utilization ratio(containers-utilization / 
> node-total-resource), P1 represents 10%~19% utilization ratio, P2 represents 
> 20%~29% utilization ratio, ..., at last P9 represents 90%~100% utilization 
> ratio. 



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to