[
https://issues.apache.org/jira/browse/KUDU-2461?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16502170#comment-16502170
]
Todd Lipcon commented on KUDU-2461:
-----------------------------------
Can think of two options:
(1) just report it as a number associated with each replica as part of the
matrix, as you suggested
(2) have some additional check whereby if any replica has not seen a healthy
leader in more than N seconds (maybe N=60 or something) then we consider it
unhealthy even if it is otherwise consistent?
I think we could certainly start with #1.
> Add election failure metrics to ksck output
> -------------------------------------------
>
> Key: KUDU-2461
> URL: https://issues.apache.org/jira/browse/KUDU-2461
> Project: Kudu
> Issue Type: Improvement
> Components: ksck, metrics, supportability
> Reporter: Attila Bukor
> Assignee: Attila Bukor
> Priority: Major
>
> KUDU-2287 added two new metrics:
> * {{failed_elections_since_stable_leader}}
> * {{time_since_last_leader_heartbeat}}
> These should be added to the ksck outputs as well
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
