On 31/08/2015 3:36 PM, Julien Danjou wrote:
On Mon, Aug 31 2015, gord chung wrote:
i'm not sure Gnocchi is where we should be fixing this as it really only
(potentially) fixes it for Gnocchi and not for any of the other ways Ceilometer
data can be consumed.
The ideal way is to send the data as it is collected and do the
treatment in the backend, otherwise you end up implementing the world in
Ceilometer.
but if it's not actually cumulative in Ceilometer (pre-storage), should
we really be tagging it as such?
a proposed solution in bug and in a previous thread suggests that a 'delta'
meter be built from cputime value libvirt provides which would better handle
the reset scenario. that said, is there another option to truly have a
cumulative meter?
Yes, the hackish way might be to convert those cumulative to delta with
a transfomer in the pipeline and be done with it. Though you'll probably
have data points missing on restart, which is not ultimate solution. As
soon as you start losing information in the Ceilometer pipeline by
transforming I'll stand and say "bad idea". :-)
so i was thinking rather than modify the existing meter, we build a new
cpu.delta meter, which is gives the delta. it seems like giving a delta
meter would make the behaviour more consistent.
So yes, you have to have a smart storage solution in the end.
ideally, i was thinking Nova would capture this data and there would be
absolutely no data loss as Nova knows exactly when instances are
restarted/stopped... that said, it's arguably outside the scope of Nova.
cheers,
--
gord
__________________________________________________________________________
OpenStack Development Mailing List (not for usage questions)
Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
