[ 
https://issues.apache.org/jira/browse/FLINK-15264?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

xiaogang zhou updated FLINK-15264:
----------------------------------
    Description: 
I am running the Single flink Job mode on Yarn. As each Job has a Job manager 
running some where in the host of the yarn. Sometimes different Job managers 
can running on the same host, and the metrics TASK_SLOTS_TOTAL has no 
identification to show which job they are belonging.

 

Can we support a metric which can tell how many slots each job occupies? like 
put a job-id description in TASK_SLOTS_TOTAL, or add another seperate metrics. 

 

[~chesnay]  seems you helped lots of people in the runtime/metrics section. let 
me know your opinion, or you can tell me any existing solution to solve the 
problem. Thanks a lot

  was:
I am run the Single flink Job mode on Yarn. As each Job has a Job manager 
running some where in the host of the yarn. 

 

Sometimes different Job managers can running on the same host, and the metrics 
TASK_SLOTS_TOTAL has no identification to show which job they belonging. Can we 
support a metric which can tell how many slot a job occupies?


> Job Manager TASK_SLOTS_TOTAL metrics does not shows the Job ID
> --------------------------------------------------------------
>
>                 Key: FLINK-15264
>                 URL: https://issues.apache.org/jira/browse/FLINK-15264
>             Project: Flink
>          Issue Type: Improvement
>          Components: Runtime / Metrics
>    Affects Versions: 1.8.1
>            Reporter: xiaogang zhou
>            Priority: Major
>
> I am running the Single flink Job mode on Yarn. As each Job has a Job manager 
> running some where in the host of the yarn. Sometimes different Job managers 
> can running on the same host, and the metrics TASK_SLOTS_TOTAL has no 
> identification to show which job they are belonging.
>  
> Can we support a metric which can tell how many slots each job occupies? like 
> put a job-id description in TASK_SLOTS_TOTAL, or add another seperate 
> metrics. 
>  
> [~chesnay]  seems you helped lots of people in the runtime/metrics section. 
> let me know your opinion, or you can tell me any existing solution to solve 
> the problem. Thanks a lot



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

Reply via email to