Re: [openstack-dev] [charms] ceilometer metering and event time-to-live defaults

[James Page]

Hi Julien

On Tue, 3 Jan 2017 at 09:17 Julien Danjou  wrote:

> > In the current ceilometer charms, we retain all ceilometer data
> > indefinitely;  the TTL can be overridden by users using configuration
> > options, but to me it feels like maybe retaining all data forever by
> > default is a trip hazard to users, and the actions required to backout
> of a
> > full DB for not that nice:
> >
> >   https://bugs.launchpad.net/charms/+source/ceilometer/+bug/1652856
> >
> > Any thoughts? What TTL defaults do deployments (and other deployment
> tools)
> > typically use for ceilometer data?
>
> The default are set to no expiry because we felt like deleting data by
> default is not a good choice. If there's a consensus that the default
> should be something else, maybe that could be fixed directly upstream.
>

Yeah - that's why we've had no expiry as a default as well.

I'm not convinced we need to switch - just felt like we could do more in a
number of ways to help users not trip over this.


> Now, I'd like to emphasize that this part of Ceilometer is officially
> deprecated, so I hope not too much effort is put into this, and more in
> the new projects that are now recommended (i.e. Gnocchi). :-)


Indeed - however the charms support older versions of OpenStack as well, so
to an extent we have to look back to the past as well as forward to the
future!

Cheers

James
__
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

Re: [openstack-dev] [charms] ceilometer metering and event time-to-live defaults

[Julien Danjou]

On Tue, Jan 03 2017, James Page wrote:

Hi James,

> In the current ceilometer charms, we retain all ceilometer data
> indefinitely;  the TTL can be overridden by users using configuration
> options, but to me it feels like maybe retaining all data forever by
> default is a trip hazard to users, and the actions required to backout of a
> full DB for not that nice:
>
>   https://bugs.launchpad.net/charms/+source/ceilometer/+bug/1652856
>
> Any thoughts? What TTL defaults do deployments (and other deployment tools)
> typically use for ceilometer data?

The default are set to no expiry because we felt like deleting data by
default is not a good choice. If there's a consensus that the default
should be something else, maybe that could be fixed directly upstream.

Now, I'd like to emphasize that this part of Ceilometer is officially
deprecated, so I hope not too much effort is put into this, and more in
the new projects that are now recommended (i.e. Gnocchi). :-)

Cheers,
-- 
Julien Danjou
/* Free Software hacker
   https://julien.danjou.info */


signature.asc
Description: PGP signature
__
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

[openstack-dev] [charms] ceilometer metering and event time-to-live defaults

[James Page]

Hi All

In the current ceilometer charms, we retain all ceilometer data
indefinitely;  the TTL can be overridden by users using configuration
options, but to me it feels like maybe retaining all data forever by
default is a trip hazard to users, and the actions required to backout of a
full DB for not that nice:

  https://bugs.launchpad.net/charms/+source/ceilometer/+bug/1652856

Any thoughts? What TTL defaults do deployments (and other deployment tools)
typically use for ceilometer data?

Cheers

James
__
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