[
https://issues.apache.org/jira/browse/HBASE-16549?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15946256#comment-15946256
]
Umesh Agashe edited comment on HBASE-16549 at 3/28/17 11:58 PM:
----------------------------------------------------------------
I have summarized AMv2 metric changes in the doc:
https://docs.google.com/document/d/1gJsjlacYRuFCSrp8q1zNRr8oWrel339rUtdmFuV7PxE/edit?usp=sharing.
Grouping related details like batch size for remote dispatch etc should not be
exposed through metrics. We can add TRACE level log messages for profiling and
debugging.
was (Author: uagashe):
I have summarized AMv2 metric changes in the doc:
https://docs.google.com/document/d/1gJsjlacYRuFCSrp8q1zNRr8oWrel339rUtdmFuV7PxE/edit?usp=sharing.
Grouping related details like batch size for remote dispatch etc should not be
exposed through metrics. We can add trace() messages for profiling and
debugging.
> Procedure v2 - Add new AM metrics
> ---------------------------------
>
> Key: HBASE-16549
> URL: https://issues.apache.org/jira/browse/HBASE-16549
> Project: HBase
> Issue Type: Sub-task
> Components: proc-v2, Region Assignment
> Affects Versions: 2.0.0
> Reporter: Matteo Bertozzi
> Assignee: Umesh Agashe
> Fix For: 2.0.0
>
>
> With the new AM we can add a bunch of metrics
> - assign/unassign time
> - server crash time
> - grouping related metrics? (how many batch we do, and similar?)
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
