14:00:46 <rafaelweingartne> #startmeeting cloudkitty
14:00:46 <openstack> Meeting started Mon May 17 14:00:46 2021 UTC and is due to finish in 60 minutes.  The chair is rafaelweingartne. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:00:47 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:00:49 <openstack> The meeting name has been set to 'cloudkitty'
14:00:52 <rafaelweingartne> Roll count
14:01:30 <priteau> o/
14:01:42 <mkarpiarz> Hi!
14:02:35 <jonher> o/
14:04:20 <rafaelweingartne> #topic Xena release
14:05:22 <rafaelweingartne> I guess that we are moving towards Xena releases. I think that it might be good to apply the agreement we had during the PTG meeting of merging things sooner than latter, and from the middle/end of the release cycle we can then focus on stabilization
14:05:44 <rafaelweingartne> having said that, I would like to ask you guys to review some of the patches I have open, they are interesting for Xena release, I guess
14:06:00 <rafaelweingartne> also, I have opened a specification for the reprocessing API that we also discussed in the PTG
14:06:13 <rafaelweingartne> it would be important to get some feedback before I move on with the implementation
14:07:07 <priteau> Agreed. If you can provide links and/or tag them as Review-Priority+1, I will allocate time for review
14:07:28 <rafaelweingartne> sure I will tag them
14:08:59 <rafaelweingartne> https://review.opendev.org/q/owner:rafael%2540apache.org+project:openstack/cloudkitty+status:open
14:09:14 <priteau> Thank you
14:09:43 <priteau> I've been a bit distracted from reviews, but for a good reason. I was working on a new cloudkitty collector
14:09:43 <rafaelweingartne> and, this one
14:09:44 <rafaelweingartne> https://review.opendev.org/q/owner:rafael%2540apache.org+project:openstack/cloudkitty-specs+status:open
14:09:51 <mkarpiarz> Sure, I should be able to help with these as well.
14:09:53 <priteau> I will try to contribute it soon
14:11:03 <rafaelweingartne> BTW, do you guys know how to add an extra style in the pdf-docs job?
14:11:15 <rafaelweingartne> When I run the job to build the pdf-docs in my laptop, it works
14:11:29 <rafaelweingartne> and it works, because I added/installed one of the latex dependencies manually
14:11:39 <rafaelweingartne> however, in the CI jobs this style does not seem to exist
14:12:10 <priteau> Do you really need to? I am not sure people really look at pdf docs
14:13:16 <rafaelweingartne> me neither... when I tested for instance, I only built the HTML ones
14:13:23 <rafaelweingartne> but the CI jobs are runing the PDF build
14:13:34 <rafaelweingartne> and this causes my specs proposal to receive a -1 from Zuul
14:14:20 <mkarpiarz> Do we have the ability to disable PDF builds then?
14:14:53 <priteau> I contributed the pdf-docs fix
14:14:56 <priteau> https://review.opendev.org/c/openstack/cloudkitty-specs/+/791624
14:15:06 <priteau> mkarpiarz: could you please review it?
14:15:19 <priteau> It's an OpenStack-wide issue, the fix comes from ironic-specs
14:16:00 <priteau> rafaelweingartne: It you rebase on top of d1803c94ef2ecb616d7bbb153eaa29bb2e51c6cd it should work
14:17:16 <rafaelweingartne> I see
14:17:28 <rafaelweingartne> I will do so then
14:17:31 <rafaelweingartne> thanks for the tips
14:22:52 <openstackgerrit> Rafael Weingartner proposed openstack/cloudkitty-specs master: Propose reprocessing API  https://review.opendev.org/c/openstack/cloudkitty-specs/+/791245
14:23:56 <openstackgerrit> Rafael Weingartner proposed openstack/cloudkitty-specs master: Propose reprocessing API  https://review.opendev.org/c/openstack/cloudkitty-specs/+/791245
14:24:51 <rafaelweingartne> done
14:25:06 <openstackgerrit> Merged openstack/cloudkitty-specs master: Update tox config  https://review.opendev.org/c/openstack/cloudkitty-specs/+/791624
14:26:45 <rafaelweingartne> Well, from my side those are the priorities
14:26:49 <rafaelweingartne> do you guys have some other?
14:26:55 <openstackgerrit> Pierre Riteau proposed openstack/cloudkitty-specs master: [ussuri][goal] Update contributor documentation  https://review.opendev.org/c/openstack/cloudkitty-specs/+/791320
14:27:39 <mkarpiarz> None at the moment.
14:28:25 <priteau> As mentioned, I have written a new collector, which uses the Nova simple tenant usage API. I will need to write some tests before submitting it
14:28:40 <rafaelweingartne> hmm
14:28:47 <rafaelweingartne> this API is scheduled or planned to be removed
14:28:48 <rafaelweingartne> right?
14:29:15 <rafaelweingartne> it also only shows for how long a VM exist in the environment, and not the actual runtime
14:29:55 <priteau> Is it planned to be removed? I've not heard about it
14:30:17 <priteau> And yes, it has shortcomings, so it isn't a replacement for Gnocchi / Monasca / Prometheus
14:30:41 <rafaelweingartne> well
14:30:41 <priteau> However, it allows you to gather information about past usage data which isn't possible if you are missing metrics from the above systems
14:30:46 <rafaelweingartne> that is what people told me when I proposed a change in the API
14:31:07 <rafaelweingartne> to allow operators to configure the API to output more meaningful data
14:31:22 <rafaelweingartne> priteau, exactly the same issue we had :)
14:31:51 <rafaelweingartne> we thought about using it, but then, when we discovered that it only shows for how long a VM exists, it was not useful anymore
14:32:20 <priteau> Do you have a link to your proposed API change?
14:32:22 <rafaelweingartne> #link https://review.opendev.org/c/openstack/nova/+/711113
14:32:28 <rafaelweingartne> you can see the discussion there
14:33:25 <priteau> Thanks
14:35:48 <priteau> It was useful for us as we believe started and stopped instances should be billed in the same way. The main issue was shelved instances being included.
14:36:03 <rafaelweingartne> ah, I see
14:36:19 <rafaelweingartne> for the cases we have applied Openstack so far that is not a valid use case
14:36:26 <rafaelweingartne> users only want to pay for actual usave of running VMs
14:36:36 <rafaelweingartne> only volumes are billed no matter if the VM is running or not :)
14:37:00 <priteau> Shut down instances consume the same amount of CPU / memory / disk resources as running ones ;-)
14:37:25 <rafaelweingartne> well, they are down
14:37:31 <rafaelweingartne> so they are not running
14:37:44 <rafaelweingartne> I guess it depends on how one handle this situations
14:38:02 <priteau> Yes but from a nova scheduler / placement point of view, they will still consume resources
14:38:08 <rafaelweingartne> it depends on the perspective public cloud versus private cloud
14:38:21 <rafaelweingartne> they are allocated, which means logically create in the system
14:38:30 <rafaelweingartne> but not actually running in the hypervisor
14:39:26 <priteau> Anyway, it's a small collector, not much maintenance, so I don't think it would hurt to include it
14:39:36 <rafaelweingartne> but anyways, I just mentioned that because we tried to used this API for a DR, and in the end, it was useless for our context
14:40:03 <rafaelweingartne> the more we aggregate the better, as long as we have a clear documentation
14:40:07 <rafaelweingartne> which was the case in Nova
14:40:27 <rafaelweingartne> people did not want to change this behavior of the API, but I suggested them, to at least clarify the behavior of the API
14:44:31 <rafaelweingartne> Well, do we have some other priority for Xena?
14:45:28 <priteau> Nothing from me
14:45:42 <rafaelweingartne> I will open for general questions then
14:45:42 <rafaelweingartne> #topic AOB
14:49:20 <priteau> rafaelweingartne: please review https://review.opendev.org/c/openstack/cloudkitty/+/789727 (trivial)
14:51:40 <rafaelweingartne> well, we are running of time
14:51:47 <rafaelweingartne> and I guess that we covered most of the planned topics
14:51:59 <rafaelweingartne> if you guys have nothing else, I will close for today
14:52:04 <rafaelweingartne> Thank you guys for participating. Have a nice week.
14:52:13 <rafaelweingartne> #endmeeting