You are viewing a plain text version of this content. The canonical link for it is here.
Posted to jira@kafka.apache.org by "John Roesler (Jira)" <ji...@apache.org> on 2020/04/30 16:52:01 UTC

[jira] [Commented] (KAFKA-7275) Prototype lock-free metrics

    [ https://issues.apache.org/jira/browse/KAFKA-7275?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17096738#comment-17096738 ] 

John Roesler commented on KAFKA-7275:
-------------------------------------

For at least one example of a higher performance, but otherwise very similar, metrics lib, see Yammer/Dropwizard metrics.

> Prototype lock-free metrics
> ---------------------------
>
>                 Key: KAFKA-7275
>                 URL: https://issues.apache.org/jira/browse/KAFKA-7275
>             Project: Kafka
>          Issue Type: Improvement
>          Components: metrics, streams
>            Reporter: John Roesler
>            Assignee: John Roesler
>            Priority: Major
>
> Currently, we have to be a little conservative in how granularly we measure things to avoid heavy synchronization costs in the metrics.
> It should be possible to refactor the thread-safe implementation to use volatile and java.util.concurrent.atomic instead and realize a pretty large performance improvement.
> However, before investing too much time in it, we should run some benchmarks to gauge how much improvement we can expect.
> I'd propose to run the benchmarks on trunk with debug turned on, and then to just remove all synchronization and run again to get an upper-bound performance improvement.
> If the results are promising, we can start prototyping a lock-free implementation.



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