14:01:19 <rafaelweingartner> #startmeeting cloudkitty
14:01:19 <opendevmeet> Meeting started Mon Mar 20 14:01:19 2023 UTC and is due to finish in 60 minutes.  The chair is rafaelweingartner. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:01:19 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:01:19 <opendevmeet> The meeting name has been set to 'cloudkitty'
14:01:26 <rafaelweingartner> Roll count!
14:01:28 <rafaelweingartner> \O
14:02:15 <mkarpiarz> Hi!
14:02:46 <priteau> o/
14:02:52 <priteau> Hello guys.
14:04:11 <rafaelweingartner> #topic vPTG  for B release?
14:04:29 <rafaelweingartner> Guys, I guess the vPTG is comming. Did we define a date and time?
14:04:54 <rafaelweingartner> I remember that we are registered for the event, but I do not recall having discussed that with you guys
14:05:00 <rafaelweingartner> what do you think?
14:06:13 <priteau> Same time Monday next week?
14:06:29 <priteau> I don't think we agreed on a time
14:06:38 <mkarpiarz> Works for me.
14:06:57 <priteau> I have a conflict with Kolla as usual, but I can skip part of it.
14:08:03 <mkarpiarz> Maybe one hour earlier or later then?
14:08:05 <priteau> Available time slots are listed at https://ptg.opendev.org/ptg.html
14:08:18 <priteau> Starts at 13UTC
14:08:28 <priteau> Remember that clocks change this weekend in Europe
14:08:45 <priteau> So 13UTC will be 14:00 in UK and 15:00 in France.
14:09:16 <rafaelweingartner> 13UTC would work for me
14:09:25 <rafaelweingartner> Ah no, next week would be a problem
14:09:37 <priteau> Is all of Monday a problem?
14:09:57 <rafaelweingartner> Next week for me, the best time would be on Tuesday
14:10:02 <rafaelweingartner> how does that work for you guys?
14:11:19 <priteau> Tuesday will be difficult
14:11:56 <rafaelweingartner> What other days work for you next week?
14:12:06 <rafaelweingartner> Next monday would be a problem to me
14:13:24 <priteau> Scheduling is going to be difficult…
14:13:37 <priteau> I have one hour Friday from 13:30 UTC
14:13:59 <rafaelweingartner> Friday would work to me
14:14:03 <rafaelweingartner> let me see
14:14:26 <rafaelweingartner> 13:30 UTC would work to me on Friday
14:15:17 <priteau> I don't think you can register on 30 minutes so you may need to book from 13UTC to 15 UTC
14:15:35 <rafaelweingartner> I think so
14:15:42 <rafaelweingartner> but we can get those slots on Friday then?
14:15:58 <priteau> Sure, why not?
14:16:03 <priteau> There's plenty of room
14:16:51 <priteau> I can book if you want
14:16:59 <priteau> It's done via IRC nowadays
14:17:03 <rafaelweingartner> yes
14:17:12 <rafaelweingartner> I am not that versed in IRC... :(
14:17:25 <rafaelweingartner> if you can, that would be much appreciated
14:18:08 <priteau> I've booked bexar room from 13 to 15 UTC on Friday
14:18:19 <priteau> Now visible here (Friday tab): https://ptg.opendev.org/ptg.html
14:18:20 <rafaelweingartner> ok, thanks
14:18:55 <priteau> So you know for next time: you just have to join #openinfra-events and run the #cloudkitty book SLOTREF (where SLOTREF is e.g. bexar-FriB1 or bexar-FriB2)
14:19:05 <rafaelweingartner> thanks
14:19:25 <priteau> We should also start an Etherpad
14:19:42 <priteau> The auto-generated one is https://etherpad.opendev.org/p/march2023-ptg-cloudkitty
14:19:49 <priteau> We can change the URL if we want
14:19:54 <rafaelweingartner> I guess we can use it
14:20:00 <rafaelweingartner> I have updated the URL in the past
14:20:05 <rafaelweingartner> but the auto generated one is fine, I guess
14:20:07 <priteau> Last one: https://etherpad.opendev.org/p/oct2022-ptg-cloudkitty
14:20:42 <rafaelweingartner> I guess we can discuss the OpenSearch and Monasca issues in the vPTG then
14:20:46 <mkarpiarz> Cool!
14:20:58 <mkarpiarz> Thanks, Pierre!
14:22:54 <rafaelweingartner> So, moving on to the reviews
14:23:02 <rafaelweingartner> #topic Target reviews
14:23:18 <rafaelweingartner> #link https://review.opendev.org/c/openstack/cloudkitty/+/861786
14:23:28 <rafaelweingartner> I created the Gabbi tests that were requested
14:23:35 <rafaelweingartner> therefore, everything should be fine now
14:23:41 <rafaelweingartner> Can you guys check it?
14:24:16 <priteau> I will look this week.
14:25:24 <rafaelweingartner> thanks
14:26:00 <rafaelweingartner> #link https://review.opendev.org/c/openstack/cloudkitty/+/861908, this one is still missing tempest tests. However, we are going to do together when we push the start/end date for rating rules that will also require tempest tests
14:26:06 <rafaelweingartner> therefore, I put that on hold for now
14:26:43 <rafaelweingartner> #link https://review.opendev.org/c/openstack/cloudkitty/+/861806, seems ready for merging. We only need a final say from priteau.
14:27:02 <rafaelweingartner> The same one goes to #link: https://review.opendev.org/c/openstack/cloudkitty/+/865417
14:29:24 <priteau> Regarding multiple rating types: I actually had a need for something like this in prometheus just last week. Do you think it could be extended to all collectors?
14:29:33 <priteau> (not necessarily in this patch)
14:29:36 <rafaelweingartner> yes, it can
14:29:41 <rafaelweingartner> we actually prepared for that
14:29:49 <rafaelweingartner> we only did not do for prometheus because we did not have the use case
14:31:40 <priteau> The only thing that I am uneasy about in this patch is the use of the term "rating types"
14:31:54 <priteau> Extends the Gnocchi collector to allow operators to create multiple rating types for the same metric in Gnocchi.
14:32:07 <rafaelweingartner> yes
14:32:14 <priteau> This term doesn't seem to be used at the moment?
14:32:14 <rafaelweingartner> that is how we express ourselves
14:32:32 <rafaelweingartner> not formally in CloudKitty, but that is what those configurations are you know
14:32:35 <priteau> I grep for "rating type" and got no result
14:32:52 <rafaelweingartner> we have a metric, and then for a metric we can create a rating, which we normally provide an alias
14:32:55 <priteau> But it may be difficult to understand for readers of the notes
14:33:10 <rafaelweingartner> so, we started using this term "rating type", so we do not confuse people
14:33:18 <priteau> Shouldn't we call them "hashmap services"/
14:33:25 <priteau> I think that's how they are used?
14:33:33 <priteau> I am not sure about Python Scripts
14:33:50 <rafaelweingartner> hashmap services are actually very confusing
14:33:54 <rafaelweingartner> at least for us
14:34:05 <priteau> But with hashmap it would be "Extends the Gnocchi collector to allow operators to create multiple hashmap services for the same metric in Gnocchi."
14:34:18 <rafaelweingartner> the whole hashmap thing was badly expressed
14:34:26 <rafaelweingartner> hashmap is just a rule that matches attributes
14:34:33 <rafaelweingartner> to activate rating rules
14:35:01 <rafaelweingartner> The whole hashmap namming seems like a fancy namming to something that is simpler when expressed in other terms
14:37:04 <rafaelweingartner> for instance, "create multiple hashmap services for the same metric in Gnocchi" does not mean much for people outside of CloudKitty world. However, "create rating multiple types for the same metric " would be (in our experience) a bit more easy for people to understand.
14:37:26 <rafaelweingartner> In the second alternative, people see right away that they are creating new types of charging for a metric
14:37:36 <rafaelweingartner> and that in CloudKitty we have a rating type that is assigned to a metric
14:37:47 <priteau> mkarpiarz: what do you think? I think these notes are for CloudKitty operators, so they should know the terminology we use
14:38:06 <rafaelweingartner> that gets more confusing with PyScript as they do not translate well to Hashmap services
14:38:39 <rafaelweingartner> as we create a pyScript for everything, and the PyScript would process all rating types
14:38:54 <rafaelweingartner> It might be a good discussion for us to have in the PTG
14:39:45 <priteau> Seems that pyscripts has a concept of services too: https://docs.openstack.org/cloudkitty/latest/user/rating/pyscripts.html#price-calculation-functions
14:40:42 <rafaelweingartner> so it would be multiple services calculations for Gnocchi metric?
14:40:45 <rafaelweingartner> or something like that?
14:41:03 <priteau> I don't know, we would have to try it with your patch
14:41:11 <priteau> Should we postpone discussion to next week?
14:41:16 <rafaelweingartner> I guess so
14:43:29 <rafaelweingartner> besides these two patches that I mentioned, we have #linl https://review.opendev.org/c/openstack/cloudkitty/+/872035
14:43:40 <rafaelweingartner> which is an optimization for Gnocchi fetcher
14:43:49 <rafaelweingartner> it would be interesting to have that patch merged soon for B release
14:46:23 <priteau> I will also try to review this week
14:47:01 <rafaelweingartner> There are also some minor improvements that would be interesting to get merged #link https://review.opendev.org/c/openstack/cloudkitty/+/877525, #link https://review.opendev.org/c/openstack/cloudkitty/+/877829, #link https://review.opendev.org/c/openstack/cloudkitty/+/877537, and #link https://review.opendev.org/c/openstack/cloudkitty/+/875753
14:47:02 <mkarpiarz> Again, this is not something I could test at scale but can look into the code.
14:51:07 <opendevreview> Rafael Weingartner proposed openstack/cloudkitty master: Optimize Gnocchi fetcher processing time  https://review.opendev.org/c/openstack/cloudkitty/+/872035
14:55:23 <rafaelweingartner> Guys we are running out of time. Do you have something else to add?
14:55:39 <mkarpiarz> Nope
14:56:33 <priteau> Let's add discussion items to the Etherpad for next week
14:56:36 <priteau> That's all from me
14:56:38 <rafaelweingartner> ok
14:56:39 <rafaelweingartner> thanks!
14:57:11 <rafaelweingartner> Thank you guys for participating. Have a nice week.
14:57:15 <rafaelweingartner> #endmeeting