[Jenkins-infra] Update 'Confluence is choking'

[Arnaud Héritier]

+1000

Clearly this monitoring isn't really efficient and nowadays the real issue
with Confluence is that it dies regularly with an OutOfMemory Exception


On Mon, Mar 27, 2017 at 12:08 PM, Olblak <me at olblak.com> wrote:

> Hello,
>
> I would like to modify 'confluence is choking' datadog warning as it
> doesn't totally reflect health status.
> Before doing that I want to be sure that nobody rely on it.
>
> Currently this check use the number of apache busy workers and trigger a
> warning if it's greater that 145.
> which doesn't mean that something bad is happening but only that traffic
> is increasing.
>
> What can goes wrong is when we reach top workers limit (250 with current
> config).
> All new Http requests are queued until a worker is available to process
> the request.
> Meanwhile Response time will increase for queued requests.
>
> It's hard to increase workers limit at the moment as more workers mean
> more cpu/memory usage
> and the host is already running out of them.
>
> I also suggest to reduce 'Confluence is slow' warning from 3sec to
> 0.1sec in order to  capture anomalies
>
> --
>   Olblak
> _______________________________________________
> Jenkins-infra mailing list
> Jenkins-infra at lists.jenkins-ci.org
> http://lists.jenkins-ci.org/mailman/listinfo/jenkins-infra
>



-- 
-----
Arnaud Héritier
http://aheritier.net
Mail/GTalk: aheritier AT gmail DOT com
Twitter/Skype : aheritier
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.jenkins-ci.org/pipermail/jenkins-infra/attachments/20170327/e47fb240/attachment.html>