12:00:01 #startmeeting watcher 12:00:01 Meeting started Thu Dec 12 12:00:01 2024 UTC and is due to finish in 60 minutes. The chair is marios. Information about MeetBot at http://wiki.debian.org/MeetBot. 12:00:01 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 12:00:01 The meeting name has been set to 'watcher' 12:00:11 hello all o/ who is around today 12:00:16 o/ 12:00:16 o/ 12:00:20 o/ 12:01:08 o/ 12:01:14 #link https://etherpad.opendev.org/p/openstack-watcher-irc-meeting meeting agenda 12:01:39 ok lets get going on the agenda items we will take them in order given 12:01:49 #topic ( rlandy) Open reviews (not expected to be merged before year end): 12:01:57 #link https://review.opendev.org/q/project:openstack/watcher+status:open+-is:wip 12:02:05 thanks @marios 12:02:07 lets add dashboard there too 12:02:17 listed the reviews there that would need W+ 12:02:18 #link https://review.opendev.org/q/project:openstack/watcher-dashboard+status:open+-is:wip 12:02:44 ok lets take them one at a time, give folks chance for any concerns or comments or status 12:02:57 #link https://review.opendev.org/c/openstack/watcher/+/934423 Add prometheus data source for watcher decision engine 12:03:12 thanks for reviews on this one i think it is close now. 12:03:20 thanks amoralej and sean for the latest round of reviews 12:03:46 my plan is to add in the client auth options and try to get it into mergeable state by the end of th eyear 12:03:51 (i.e. next week ;)) 12:04:03 then the followup patch will deal with vm level metrics 12:04:11 you mean auth or better https handling ? 12:04:18 i mean both 12:04:30 i thought we wouldn't need auth 12:04:33 basicaaly add the client opts for basic_auth user/pass if user wants, or client cert/ca 12:04:39 from our side its just adding the client opts 12:04:52 the "heavy lifting " is done on the observabilityclient 12:04:59 does observabilityclient support auth ? 12:05:02 so agree with the notion of trying to establish the interface 12:05:07 it has basic_auth 12:05:12 you can set the user/pass on the session 12:05:15 if you pass it through 12:05:16 good 12:05:21 yeah, then, let's expose it 12:05:28 so it should be cheap to add the client opts on our side anyway 12:05:28 but in follow up :) 12:05:32 and establish our interface 12:05:45 ok lets discuss this bit on the review then, personally i think one more iteration for th eclient opts is OK 12:06:12 ok 12:06:26 thanks amoralej for diving in there and as discussed earlier in this channel helping to test this it is very useful and helpful feedback 12:06:52 thanks to others that have had a look already and please keep reviews coming thank you 12:06:57 anything else before we move on from this one 12:07:41 #link https://review.opendev.org/c/openstack/watcher/+/935371 Add Tempest test for Prometheus integration 12:07:50 thanks for all the reviews 12:07:59 only W needed here 12:08:14 nice one rlandy++ 12:08:22 working on scenarios in follow up patch 12:08:40 but hitting the stated bug - so will report on that in next meeting 12:09:03 so this patch is setting up sg-core to expose ceilometer metrics exported as prometheus metrics 12:09:20 i mean 935371 (adding some general info for anyone following along in the meeting logs) 12:09:42 but currently it is actually not using any of the prometheus metrics right? 12:09:47 so that is coming in the followup patch 12:09:55 correct 12:09:59 lets link it here? 12:10:20 https://review.opendev.org/c/openstack/watcher/+/937497 12:10:43 #link https://review.opendev.org/c/openstack/watcher/+/937497 WIP: Run scenario tests with prometheus 12:11:05 anything further on this topic rlandy or anyone? 12:11:06 you will notice they all fail on https://bugs.launchpad.net/watcher-tempest-plugin/+bug/2090853 12:11:21 so we have no option but to get the prometheus metrics in :) 12:11:26 that's all 12:12:13 #info known bug on watcher-tempest-plugin affecting https://bugs.launchpad.net/watcher-tempest-plugin/+bug/2090853 12:12:28 thanks for the info rlandy 12:12:37 anything else here before we move to the next 12:13:02 #link https://review.opendev.org/c/openstack/watcher/+/933909 Update python versions, drop py3.8 12:13:33 looks like martin is not here right now? 12:13:54 anyone else can talk about this one? i think as written, we just need a workflow . 12:14:06 our only current active core has already added +2 (Sean) 12:14:18 marios: note these were just added to the list to notify people that we are waiting on W ... idk if there is a need to address any one in particular 12:14:28 martin is out until monday 12:14:45 so right now it would mean harrassing dsmith or slaweq as the other two cores. 12:14:55 probably we should add them to the review, doing that now in fact with a note 12:16:14 Dan, Slaweq can you please check this and merge if you agree. Right now we only have Sean as active core so until we get some more please help us merge some of the things. Sorry for unsolicited pings and thanks in advance ;) 12:16:19 i added this note ^ 12:16:23 anything further on this topic? 12:16:41 i think if there is no push back until the new year we can ask sean to merge like this 12:16:48 that is another solution until we build our core team 12:16:58 +1 12:17:13 ie: ok to wait 12:17:39 yes 12:17:43 ok, moving on ... 12:17:52 #link https://review.opendev.org/c/openstack/watcher/+/934583 Remove default override for config options policy_file 12:18:08 i think similar story here right 12:18:27 this one belongs to gmann 12:19:03 so sean has voted and we'll need to harrass Dan/Chenker/Slaweq for now, and in the new year we can ask sean to merge with one +2 if needed until we build core team out 12:19:42 i won't copy/paste my review request note on this review (or others) yet for example i think some folks will indeed be out until the new year 12:19:52 anything else on this one before we move on? 12:20:16 #link https://review.opendev.org/c/openstack/watcher/+/932264 tox: Drop envdir 12:20:33 same story here ... 12:20:47 ah in fact, here sean can merge it 12:20:56 dsmith added his +2 12:21:08 adding note for sean 12:21:45 anything else here? 12:22:05 #link https://review.opendev.org/c/openstack/watcher/+/934996 Unblock the CI gate 12:22:53 so per seans comment here we need to wait for https://review.opendev.org/c/openstack/project-config/+/936754 12:23:08 but then we should workflow -1 /934996 or add depends-on 12:24:27 moving on ... 12:24:37 next topic 12:24:45 #topic (amoralej) For watcher-dashboard the more meaningful reviews are: 12:24:58 amoralej: i will list them here but feel free to start giving any context you wanted to 12:25:08 #link https://review.opendev.org/c/openstack/watcher-dashboard/+/932693 12:25:14 #link https://review.opendev.org/c/openstack/watcher-dashboard/+/888912 12:25:19 #link https://review.opendev.org/c/openstack/watcher-dashboard/+/933910 12:25:26 #link https://review.opendev.org/c/openstack/watcher-dashboard/+/935043 12:25:35 #link https://review.opendev.org/c/openstack/watcher-dashboard/+/934781 12:25:37 we don't need to go one by one, i just wanted to focus what are the more relevant reviews in watcher-dashboards 12:25:44 #link https://review.opendev.org/c/openstack/watcher-dashboard/+/932890 12:25:56 ack anything we need to call out? 12:26:01 it's mostly some sanity fixes 12:26:23 and bug fixing 12:26:57 just, asking for review and we can discuss there or here if you have doubts 12:27:31 thanks amoralej any comments before we move on 12:27:36 (anyone) 12:28:13 I was wondering if we should backport those fixes in watcher-dashboard to supported releases 12:28:41 I'm not sure if it's worthy, tbh 12:29:07 so, it might be more work than worth it, but i am also not sure yet 12:29:25 the project has been mostly dormant in the last couple years so ... might be easy/cheap to backport 12:29:30 but does anyone care? 12:29:48 it seems there is enough work to get them into the current release 12:29:54 I expect backports should be easy to do, but yes, that's my question 12:30:01 if anyone is interested 12:30:15 and it's mostly work to get ci running, etc... 12:30:21 in case it's not, i didn't check 12:30:37 yes this is another good point rlandy ; can we revisit this in the new year amoralej sounds like a good agenda item once these things actually start mergein on master 12:30:39 we can focus in master so far 12:30:53 +1 12:31:30 k moving on .. 12:31:37 +1 on revisit this, once we have more reviewers too 12:31:54 reviewers that have power to merge them :) 12:32:04 yes 12:32:07 +1 12:32:28 #topic https://review.opendev.org/q/project:openstack/watcher-specs+status:open+-is:wip 12:32:37 #link https://review.opendev.org/c/openstack/watcher-specs/+/752508 Fix hacking min version to 3.0.1 12:33:17 i *think* takashi had a series of patches for this no? 12:33:19 checking 12:33:35 ah this is watcher-specs though maybe he missed that one 12:34:13 k cant quickly find anyway 12:34:23 any discussion on this one? 12:35:04 moving on momentarily .. 12:35:32 #topic https://review.opendev.org/q/project:openstack/watcher-tempest-plugin+status:open+-is:wip 12:35:37 #link https://review.opendev.org/c/openstack/watcher-tempest-plugin/+/936837 Update documentation and add a job to publish it 12:35:48 #link https://review.opendev.org/c/openstack/watcher-tempest-plugin/+/934009 Drop translation sections 12:35:56 any dicussion here ? 12:36:49 moving on in a moment... 12:37:35 #topic (doug) Meetings on Dec 26th and Jan 2nd. Should we cancel these due to holidays? 12:37:55 thanks for adding this one i was going to propose the same 12:38:00 yep, due to holidays and low audience, is anyone against to cancel both? 12:38:01 looks like there is agreement on the etherpad 12:38:35 +1 to cancel, there's likely very few people around on those dates 12:39:17 #info watcher weekly irc meeting cancelled Thursday 26 December 2024 and Thursday 02 January 2025 due to engineering holiday outages 12:39:36 anything else before we move on ? 12:39:54 we need someone to send that in the ML 12:40:01 yes good call dviroel 12:40:05 any volunteer 12:40:20 i can 12:40:48 #action dviroel will mail openstack-discuss about cancelled irc meetings 26Dec & 02Jan 12:40:51 thank you 12:41:32 #topic (doug) Nova spec merged: scheduler hints info available in server details 12:41:38 #link https://review.opendev.org/c/openstack/nova-specs/+/936140 Add spec for show scheduler hints in server details 12:41:49 yeap, just want to bring this topic, since is important from watcher's point of view 12:42:27 nice job getting that merged dviroel 12:42:28 this spec is now merged in nova-spec, and one of the main use cases is to allow services, like watcher to retrieve additional placement constraints of running instances 12:42:30 nice 12:42:39 dviroel++ 12:42:52 for watcher, this could reflect on more constraints to be evaluated by Strategies, when selecting a destination host for instance migrations 12:43:15 I am expecting to propose the nova changes and get them merged in this release too 12:43:42 and after that, watcher and strategies will also need to be modified to properly account these new constraints... 12:43:58 but probably work for another release 12:44:55 any questions on that $topic? 12:45:05 thanks for the extra context dviroel 12:45:08 you think some initial improvements in the strategy to align with nova scheduler, other that the related with the scheduler-hints, is doable in this release? 12:45:09 :) 12:46:06 amoralej: maybe, it is also possible to include other constraints like server_groups for instance... 12:46:18 that are already available in nova's api 12:46:18 exactly, that was my point 12:46:50 given that we already have part of it, i was thinking if we may start implementing that part which does not require the new feature 12:47:01 even if it's not perfect, it may be better that what we have 12:47:22 said so, i have no idea about how complex it is 12:48:00 right, I think that is something that I can bring to the following meetings... 12:48:14 sounds good to me 12:48:30 anything further from anyone for right now on this before we move on 12:49:29 #topic next meeting chair 12:49:47 so rlandy is in the list but i don't know if they will be up for it... we can revisit between now and then 12:49:52 I may not be available at meeting time next week 12:50:04 but i think rlandy said they can do it so leaving as is for now 12:50:05 I can do it next week 12:50:20 #info rlandy to chair irc meeting 19th December 2024 12:50:29 next one after that is (checks calendar) 12:50:38 09th Jan 12:50:56 volunteer to chair 09 Jan irc meeting please? 12:51:17 i can take it 12:51:18 if you aren't up for next week i can take it rlandy we talk before then obv 12:51:26 thank you amoralej 12:51:41 #info amoralej to chair irc meeting 09 Jan 2025 12:51:51 thanks amoralej 12:52:10 that is the end of topics 12:52:17 #topic open discussion 12:52:30 anything anyone would like to bring up here before we close out today? 12:53:57 ok then thanks all for participating o/ 12:53:59 #endmeeting