[openstack-dev] [Ceilometer] Ceilometer and nova compute cells

Doug Hellmann doug.hellmann at dreamhost.com
Fri Aug 2 11:40:29 UTC 2013


On Thu, Aug 1, 2013 at 7:36 PM, Sam Morrison <sorrison at gmail.com> wrote:

>
> On 31/07/2013, at 6:45 PM, Julien Danjou <julien at danjou.info> wrote:
>
> > On Wed, Jul 31 2013, Sam Morrison wrote:
> >
> > Hi Sam,
> >
> >> Does everything that gets stored in the datastore go through the
> >> ceilometer.collector.metering queue?
> >
> > If you only use the RPC publisher (which is the default), yes.
> >
> >> If so maybe the collector could instead of storing the message forward
> these
> >> messages to another rabbit server where another collector is running
> who in
> >> turn either forwards it up or stores it in the datastore (or both maybe)
> >> I think the confusing thing to me is how notification messages get into
> >> ceilometer. The collector does 2 things I think? And it seems to leave
> the
> >> messages on the notification.info queue to build up?
> >
> > Yes, the collector has 2 purposes right now:
> > 1. It listens for notifications on the RPC bus and converts them into
> >   Samples;
> > 2. It publishes these Samples according to your pipeline.yaml file to
> >   different conduct, the default being a RPC call on a collector for
> >   storing these samples.
>
> Thanks for the info, looking at the code I think the way to do this would
> be to create a new dispatcher that instead of recording the metering data
> just forwards it on to another RPC server for another collector to consume.
> Do you think that's the way to go?
>

I'm not certain any new code needs to be written. Couldn't we configure the
pipeline in the cell to send the data directly upstream to the central
collector, instead of having it pass through a collector in the cell?

Doug


>
> Cheers,
> Sam
>
>
>
> >
> > --
> > Julien Danjou
> > ;; Free Software hacker ; freelance consultant
> > ;; http://julien.danjou.info
>
>
> _______________________________________________
> OpenStack-dev mailing list
> OpenStack-dev at lists.openstack.org
> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack-dev/attachments/20130802/0cc14efd/attachment.html>


More information about the OpenStack-dev mailing list