zookeeper-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Damien Diederen (Jira)" <j...@apache.org>
Subject [jira] [Commented] (ZOOKEEPER-4211) Expose Quota Metrics to Prometheus
Date Thu, 06 May 2021 18:05:00 GMT

    [ https://issues.apache.org/jira/browse/ZOOKEEPER-4211?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17340353#comment-17340353

Damien Diederen commented on ZOOKEEPER-4211:

Hi [~liwang],

Ha, funny; I just caught up with your discussion on the {{dev}} mailing list!  (https://mail-archives.apache.org/mod_mbox/zookeeper-dev/202105.mbox/browser).
 Excellent work, btw.

I haven't forgotten about your points, but have just been really busy on other topics.  (Of
course, don't hesitate to ping other reviewers, particularly when I am lagging!)

That being said:

bq. Based on investigation result of the performance impact of Prometheus Summary quantile
computation, I am working on adding the support for CounterSet for the use case that need
to group counter metrics by keys (i.e. top namespace) but no need for quantiles and sum.

I figure we'll have to use some sort of queue to gather all these metrics without blocking
the worker threads, but +1 on not making things worse in the meantime :)

> Expose Quota Metrics to Prometheus
> ----------------------------------
>                 Key: ZOOKEEPER-4211
>                 URL: https://issues.apache.org/jira/browse/ZOOKEEPER-4211
>             Project: ZooKeeper
>          Issue Type: New Feature
>          Components: metric system
>    Affects Versions: 3.7.0, 3.7
>            Reporter: Li Wang
>            Priority: Major
>              Labels: pull-request-available
>          Time Spent: 5.5h
>  Remaining Estimate: 0h
> In 3.7, Quota limit can be enforced and the quota related stats are captured in the StatsTrack.
 From the "listquota" CLI command, we can the quota limit and usage info. 
> As an addition to that, we would like to collect the quota metrics and expose them to
the Prometheus for the following:
> 1. Monitoring per namespace (Chroot) quota usage via the Grafana dashboard
> 2. Creating alert based on the quota levels (e.g. 90% used)

This message was sent by Atlassian Jira

View raw message