[jira] [Commented] (CASSANDRA-10718) Group pending compactions based on table

["Tushar Agrawal (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/CASSANDRA-10718?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15034335#comment-15034335 ] 

Tushar Agrawal commented on CASSANDRA-10718:
--------------------------------------------

Need some help and guidance:

PendingTasks metrics comes from "pendingTasks" gauge in CompactionMetrics.java. In getValue() method it is currently looping through all keyspaces and tables to get the estimated remaining tasks.

Shall I change the getValue() method (and call hierarchy) to return a Map<KeyspaceName, Map<tablename, pendingTasksValue>> ? How this would be displayed under JMX metrics?

What could be the other approaches?



> Group pending compactions based on table
> ----------------------------------------
>
>                 Key: CASSANDRA-10718
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-10718
>             Project: Cassandra
>          Issue Type: Improvement
>            Reporter: Marcus Eriksson
>            Assignee: Tushar Agrawal
>            Priority: Minor
>              Labels: lhf
>             Fix For: 3.x
>
>
> Currently we only give a global number on how many compactions are pending, we should group this on a per-table basis, example:
> {code}
> $ nodetool compactionstats
> pending tasks:
> - keyspace1.standard1: 10
> - other_ks.table: 2
> {code}



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)