You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@cassandra.apache.org by "Jason Brown (JIRA)" <ji...@apache.org> on 2017/03/16 13:18:41 UTC

[jira] [Commented] (CASSANDRA-12944) Diagnostic Events

    [ https://issues.apache.org/jira/browse/CASSANDRA-12944?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15928004#comment-15928004 ] 

Jason Brown commented on CASSANDRA-12944:
-----------------------------------------

On the whole I do like this idea. Do you know if there are any other OSS systems that have a similar function?

I've added some comments to the RFC that are specific to the text there, but I'm adding some broader thoughts here.

It feels like a lot conceptually here that overlaps with tracing. In practice and implementation, I'm not sure there will end up being any overlap, but at certain call sites I can imagine the code sending out both tracing and diagnostic events. Not sure if that's avoidable or a problem, but I'm wondering if you've thought about any overlap.

Speaking of tracing, we do not encourage users to turn on tracing and leave it on for everything as they'll crush the server process with the tracing events. I think we want to avoid that kind of situation at all costs with diagnostic events, else we lose some value of this ticket. Do you imagine users to let this functionality run all the time when the database is running? (It would seem so, and I agree.)

> Diagnostic Events
> -----------------
>
>                 Key: CASSANDRA-12944
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-12944
>             Project: Cassandra
>          Issue Type: Improvement
>          Components: Observability
>            Reporter: Stefan Podkowinski
>
> I'd like to propose a new "diagnostic events" feature that would allow to observe internal Cassandra events in unit tests and from external tools via native transport. The motivation is to improve testing as well as operational monitoring and troubleshooting beyond logs and metrics.
> Please find more details in the linked proposal and give it some thoughts :)



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)