[openstack-dev] [ceilometer] Pollsters and counters

Julien Danjou julien at danjou.info
Thu Jan 17 09:12:43 UTC 2013

On Thu, Jan 17 2013, Jiang, Yunhong wrote:

> a) as discussed before, the pipeline definition should not be changed in
> deployment unless expert user. While in deployment, user may want to disable
> some counter specifically.

I don't think we are sure about that. The pipeline definition is also
going to configure the periodicity of pollsters for example. It's really
likely that it's going to be modified.

AFAIK the assupmtion that the pipeline is not going to be modified has
been asserted because it seems that the pipeline is going to be complex
for publisher like CloudWatch because of the amount of transformers
If this is a problem, this is likely to be hidden in another way, like
the ability of defining a transformer based on several transformer (some
sort of macro).

> b) If we really have several publishers (3 or 4, or even more), that means
> user have to excluded the counters in each pipeline definition and that's
> not so convenient. After all, pipeline is mostly a per-publisher
> configuration.

I don't think it's really a problem to disable counters when you have 3
or 4 publishers. People configuring 3 or 4 publishers will have to know
what they're doing.

> BTW, I noticed the glance pollster is one counter per pollster while swift
> pollster is all counters in one pollster. I think we need keep it same for
> all resource pollster.

Yes, don't worry, we know the current state is not ideal. But it's
working, that's already a good point. :)

Julien Danjou
-- Free Software hacker & freelance
-- http://julien.danjou.info
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 835 bytes
Desc: not available
URL: <http://lists.openstack.org/pipermail/openstack-dev/attachments/20130117/ee21f001/attachment.pgp>

More information about the OpenStack-dev mailing list