Thursday, 2026-04-09

opendevreviewchandan kumar proposed openstack/watcher-specs master: Add spec for improving watcher-dashboard testing  https://review.opendev.org/c/openstack/watcher-specs/+/97022006:05
opendevreviewchandan kumar proposed openstack/watcher-dashboard master: Add Playwright-based E2E testing framework  https://review.opendev.org/c/openstack/watcher-dashboard/+/97035306:39
opendevreviewchandan kumar proposed openstack/watcher-dashboard master: Add Playwright-based E2E testing framework  https://review.opendev.org/c/openstack/watcher-dashboard/+/97035306:52
opendevreviewchandan kumar proposed openstack/watcher-dashboard master: Add Playwright-based E2E testing framework  https://review.opendev.org/c/openstack/watcher-dashboard/+/97035307:30
opendevreviewchandan kumar proposed openstack/watcher-dashboard master: Add Playwright-based E2E testing framework  https://review.opendev.org/c/openstack/watcher-dashboard/+/97035308:23
opendevreviewchandan kumar proposed openstack/watcher-dashboard master: Add Playwright-based E2E testing framework  https://review.opendev.org/c/openstack/watcher-dashboard/+/97035309:13
dviroelhello all, watcher meeting will start in 8 min, please add your topics to our meeting etherpad: https://etherpad.opendev.org/p/openstack-watcher-irc-meeting#L3011:52
dviroel#startmeeting watcher12:00
opendevmeetMeeting started Thu Apr  9 12:00:55 2026 UTC and is due to finish in 60 minutes.  The chair is dviroel. Information about MeetBot at http://wiki.debian.org/MeetBot.12:00
opendevmeetUseful Commands: #action #agreed #help #info #idea #link #topic #startvote.12:00
opendevmeetThe meeting name has been set to 'watcher'12:00
amoralejo/12:01
dviroelhi all o/12:01
morenodo/12:01
jgilabero/12:01
sean-k-mooneyo/12:01
chandankumaro/12:01
dviroelcourtesy ping: 12:01
dviroelah! everybody is around :)12:02
dviroellet's start with today's meeting agenda12:02
dviroel#link https://etherpad.opendev.org/p/openstack-watcher-irc-meeting#L30 (Meeting agenda)12:02
dviroelas usual, feel free to add your own topics to the agenda12:02
dviroel#topic Any objection to moving the stable/2024.2 to eol12:03
dviroeljgilaber: thanks for raising this topic12:03
dviroelso according with12:03
dviroel#link https://releases.openstack.org/12:03
dviroelthe Dalmatian eol is estimated 2026-04-2912:04
amoralejisn't EOL coordinated to all projects after the last change in lifecycle?12:04
amoraleji mean coordinated from TC12:04
amoralejor release project, i'm not sure12:05
jgilabersupposedly, but in reality it did not happen last cycle12:05
dviroelyeah, I think that the question is if we are happy with its content or if we want to merge more backports?12:05
jgilaberI submitted a patch to move unmaintained/2023.1 to eol12:06
jgilaberso we can wait until the estimated date and see if it happens12:06
dviroelyeah, but I remember a discussion around that at that time 12:06
jgilaberbut we can also dicuss what dviroel suggests, is there anything we want to get there before eol?12:06
amoralejhow was 2023.2 managed?12:07
jgilaberI guess not but it's worth to highlight it here I think12:07
amoralejfor 2023.2 the branch was removed without intervention from the project, i guess?12:07
morenodquestion: when a version reachs eol, we remove it from gate and check jobs?12:08
sean-k-mooneyo/12:08
jgilabersean-k-mooney, sent a patch to move it to eol https://github.com/openstack/releases/commit/3a9a33528cf532be0bad1cfbfb8679742d02437412:08
amoralejyes, actually, the entire branch is removed12:08
sean-k-mooney yes 12:08
amoralejthat's just creating a new tag, not managing branches12:08
amoralejso, yeah, we should do the same12:09
sean-k-mooneyso 2024.2 is not eligable to be unmaintained12:09
sean-k-mooneyno .2 brabces are and they move to eol directly12:09
amoraleji'd expect the releases project to send the patch for the tag, but we can also do it12:09
sean-k-mooneyonly slurp release can go to unmainteied and there shoudl only ever be 1 unmainteind branch under the polciy12:09
sean-k-mooneyya elod will likely do it en mass but we can do it our selves12:10
dviroelthat's my expectation too12:10
sean-k-mooneyi have a ptach up for the cybrog-tempsest-plugin i belive to drop the 2024.2 job but we need to do the same for watcher12:11
amoralejabout the question of backports https://review.opendev.org/q/project:openstack/watcher+branch:stable/2024.2 maybe we could spend some time merging the open ones passing ci ?12:11
jgilaberso by the end of the month both 2024.1 and 2024.2 should move to eol, right?12:11
dviroelbut now about our 2024.2 branch, do we have someone available to work on getting unit tests and ci passing?12:11
sean-k-mooney2024.2 is blocked due to issue win the requriement project12:11
sean-k-mooney2024.1 wont be eol unless we chosoe not to opt back in12:11
sean-k-mooneybut it would be eol next release when 2025.1 is demoted form stable to unmaintaned12:12
sean-k-mooneycurrently the stable branchs are 2026.1 2025.12:12
sean-k-mooneycurrently the stable branchs are 2026.1 2025.2 and 2025.112:12
sean-k-mooneyso when 2026.2 is relsaed 2025.1 -> unmainteded and 2024.1 -> eol12:13
jgilaberah ok, I misread the timeline and though 2025.1 would become unmaintained now, but it's next cycle, got it12:13
amoralejyep ^ that's it12:13
sean-k-mooneyi will note that the ptl can retire unmainted barnches early at any time12:14
sean-k-mooneythere exisating requries an opt in12:14
sean-k-mooneyand if the gates are not passing we can retire them for lack of maiatnace12:14
sean-k-mooneyso if we chose to we can eol 2024.1 early since it broken because of pkg_resouces12:14
sean-k-mooneybut we dont have too12:14
sean-k-mooneypassing ci is ment to be a requiremnt at the start of each release so techinally since it not green it shoudl eb retired12:15
dviroelnot sure how reliable is 2024.1, i didn't really deployed this release12:16
sean-k-mooneyya its a judgement call at the end of the day12:16
amoraleji just recheck a patch to see how ci is failing12:16
amoralejis the pkg_resources the problem there?12:17
sean-k-mooneyamoralej: it probaly going to fail on setuptools droping pkg_resouces12:17
amoralejyeah, let's see12:17
sean-k-mooneyamoralej: yes so that is not fixed in the requireemnt repo on 2024.2 so 2024.1 is also broken as are most older relases12:17
amoralejyep https://zuul.opendev.org/t/openstack/build/b59e34afb9ea4b27a8a99c473ab25b4e12:18
sean-k-mooneybut lest see what your patch show 12:18
amoralejthat's from my recheck12:18
sean-k-mooneyfor a while the older banchces tghat ran on jammy were passign but that broke after virutalenv updated there bundeled setuptools12:19
sean-k-mooneywell thats on stable/2024.2 not stable/2024.112:19
amoralejso, i wouldn't invest any more time in the backports12:19
sean-k-mooneyright i would however say doing a 2025.1 release might be nice12:19
amoralej+112:20
sean-k-mooneyso spending some tiem on the supprot stable branches 12:20
dviroelright12:20
sean-k-mooneychandankumar: out of interst did you or jgilaber update the serise to mark 2026.1 as the newest supproted branch12:20
sean-k-mooneyand mark 2024.2 as unsupproted (obsolete?) in lachpad yet12:21
sean-k-mooneynow that the officall release has happend we shoudl update that12:21
chandankumarsean-k-mooney: thank you for the ping, let me do that12:21
sean-k-mooneyack12:22
sean-k-mooneyis there more on this topic or shall we move on12:22
dviroellets move on12:22
dviroel#topic PTG12:22
dviroelour ptg planning etherpad is available at: Call for topics: https://etherpad.opendev.org/p/watcher-2026.2-ptg12:23
dviroelso, if you plan to discuss anything in watcher, that is not cover by the already proposed topics12:23
dviroelfell free to add new ones12:24
dviroelyou can also choose to work with someone in a specific topic12:24
sean-k-mooneywe shoudl proably ahve a deprecatoion/deletion topic12:24
sean-k-mooneybasicly maas/ironic deprecaiton and removal12:25
sean-k-mooneyor some other topic like that12:25
dviroelright,  i was planning to cover in the Datasources and Integrations topicv12:25
sean-k-mooneyack that works12:25
dviroelthere is also the Prometheus there to discuss12:25
sean-k-mooneyif we dont add an ironci job this cycle i think we shoudl jsut look to remove it12:25
dviroelack, we should also have a testing/ci topic that can prioritize this12:26
sean-k-mooneybut we can discss the details and timelien in the ptg12:26
dviroelyeah12:26
dviroeldo we all agree with the initial schedule for Watcher?12:26
sean-k-mooneyi am wondering if we want to add some cybrog awareness this cycle12:26
dviroelhttps://ptg.opendev.org/ptg.html12:26
sean-k-mooneythat coudl be a topic.12:27
sean-k-mooneyi.e. you cant live migrate with cyborg so make our stagies aware of that12:27
dviroelthat could be a good topic yes12:27
sean-k-mooneydviroel: in terms of timetable i think that good12:27
opendevreviewchandan kumar proposed openstack/watcher-dashboard master: Add Playwright-based E2E testing framework  https://review.opendev.org/c/openstack/watcher-dashboard/+/97035312:27
sean-k-mooneythe fact that we start an horu earlier and end an hour eailr then nova each day give at least 1 hour a day wehre we dont have a conflict12:28
sean-k-mooneyyou also avoid the tc slots and i thinke 3 days shoudl be enouch time12:28
sean-k-mooneyi booked the cyborg slot for monday/tuesday so there will be no conflcit 12:28
dviroelwe can,  the PTG schedule track doesn't provide that slot only (the one hour earlier)12:29
sean-k-mooneywell no its already correct12:29
sean-k-mooneywather starts at 13:00 utc -> 16:00 utc12:30
sean-k-mooneynova is 14:00-17:00 slot12:30
sean-k-mooneyso both are 4 hours with a one hour offset12:30
sean-k-mooneyso that good i think12:30
dviroeloh, i understood that you were suggesting starting one hour earlier (i.e 12:00 utc)12:31
sean-k-mooneyno 12:31
dviroelah ok12:31
sean-k-mooneyi think the slots you booked are good12:31
dviroelack12:31
dviroelanything else in this topic? we should revisit this next week12:32
amoraleji'll check the list and update the scalability part12:32
dviroelfor now, please focus on adding topics to the etherpad, so we can check the sessions that you plan to cover/discuss12:32
dviroelamoralej: ack tks12:32
dviroel#topic Eventlet-removal12:33
dviroelthis is just a quick update12:33
dviroeljust to note that threading is now the default concurrency mode12:33
dviroel#link https://review.opendev.org/c/openstack/watcher/+/98081812:33
amoralejdviroel++ 12:34
dviroelthis patch merged 2 days ago, so when deploying watcher you will not be using eventlet by default anymoew12:34
amoralejso the only remaining task is to remove eventlet option in one or two releases, right?12:34
dviroeljust in case you see any bug, performance issues, please feel free to reach us or file a bug12:35
dviroelamoralej: correct, i will update the DNM patch (that removes all eventlet code) that I have to get more CI updates12:35
amoralejgreat12:35
dviroelwe should have a 10m session about evenlet removal in PTG12:36
dviroelwe can agree on the next step about the removal12:36
dviroelso, moving to next topic12:36
sean-k-mooneyso the proejct wide concensou seam to be eventlet removal does nto require deprecation12:37
sean-k-mooneywhich means we can remove it at any time bu ti woudl sugete we shoudl only do it after we have had a relase with it as the defualt12:37
dviroelyeah12:37
sean-k-mooneyso removal woudl not happen until at least 2027.112:37
dviroelack, and we will be still in the community schedule too12:38
sean-k-mooneythat also what nova is working towards i think12:38
sean-k-mooneyi think there will be an eventlet cross project at the ptg12:38
sean-k-mooneyso we shoudl try to attend that12:38
dviroelyep, it may happen early in the ptg week12:38
dviroelok, next one is from amoralej 12:39
dviroel#topic Freezeing the reply to API datamodel list12:39
amoralejin last PTG we discussed about that in order to avoid requiring to bump API microversion on model updates12:39
dviroelright12:40
amoralejfreeze the content of the api call with current state and treat it as a internal thing12:40
amoralejdid we agree on the approach?12:40
sean-k-mooneyya so it frozen by default as we are nto allwoe dot make change the alter the api respocen without a microverion12:40
amoralejseems reasonable to me12:40
dviroelwe agreed on freezing it, but there is nothing blocking it in code/testing12:41
amoralejif we agree on that, it's something we could start working on as it would help other changes in the todo list12:41
sean-k-mooneydviroel: well that woudl violate the microverson contract12:41
sean-k-mooneythis is ment ot be tested and block by the api sample tests12:41
amoralejso, to manage the freeze itself, what would we need?12:41
amoralejin terms of tracking, i mean12:41
amoralejspecless blueprint?12:42
sean-k-mooneynoting we are not changign anything12:42
sean-k-mooneythe current polciy is any change that has an api impact needs a spec12:42
amoralejbut we need code so that changes in the model are not shown in the api call12:42
sean-k-mooneyyes12:42
amoralejso that would be managed as a ... bug? feature?12:42
sean-k-mooneybut we need to review for that and -2 any change that would alter any api respoce without an aprpoved spec12:42
amoralejor nothing :) 12:43
dviroelyes, currently the code gets the model from decision-engine and expose it 12:43
amoralejyep, that's my point, we need code to filter the fields to be included if anything new appears12:43
sean-k-mooneyso as i said above we should have api sample tests for every api microverion and evnetually openapi schemas too12:43
sean-k-mooneyso both of those if present shoudl fail if we acidnetly modfiy somehting12:43
dviroelbut I guess that today our unit tests fail already if a new attribute is added12:44
sean-k-mooneyhttps://github.com/openstack/watcher/tree/master/api-ref/source/samples12:44
sean-k-mooneythose shoudl already deteact this in theory12:44
sean-k-mooneybut we need to verify that12:44
sean-k-mooneyhttps://github.com/openstack/watcher/blob/master/api-ref/source/samples/datamodel-list-response.json12:44
sean-k-mooneymy guess is we don not currently ahve complete coverage12:45
dviroelnot sure if we detect additional kv 12:45
sean-k-mooneyhttps://github.com/openstack/watcher/commit/03c09825f7421eac429034fb4e25ba089671077d12:45
sean-k-mooneywell you extended it when you last did it12:45
sean-k-mooneyso i assume that fialed12:45
sean-k-mooneywe can tst it by just addign a filed in a DNM12:46
dviroelyeah, i added specific test for different microversion, but I will need to test again to see if it fails on LATEST with new content12:46
* dviroel it should :) 12:46
dviroeli  will check that and provide a feedback12:47
amoralejso the scope of the required change would be to add fields filtering in the api call + coverage to discover if anything new appears in the api call12:47
sean-k-mooneyso in other proejct the api sample tests are the primary way we verify microverion stablity12:47
sean-k-mooneybut we layer other test on top as well12:47
amoralej^ am i correct?12:48
sean-k-mooneyamoralej: you can basilly add a test=true12:48
sean-k-mooneyfiled to hte model12:48
dviroelyeah, and see if tests will fail12:48
sean-k-mooneytaht shoudl cause the test to fail just hard code something12:48
amoralejack12:48
sean-k-mooneywhen we do it for real 12:48
dviroelbecause in the past, that was a silent api change12:49
sean-k-mooneywe will need to updat ethe respoce code to filter out the unsupproted files12:49
sean-k-mooneythe same way that doug did for the older microveriosns12:49
amoralejyes, that was my initial point12:49
sean-k-mooneyright that not a bug12:49
sean-k-mooneythat a review reuqirement for any future spec that adds new datamodel files12:49
sean-k-mooneyor any bugfix for the same12:49
amoralejyep12:50
dviroelalredy, amoralej lets sync afterwards that here in the channel12:50
amoralejok12:50
amoralejthanks12:50
dviroelthanks for raising the topic12:51
amoraleji think we can move on12:51
dviroel#topic Reviews12:51
dviroelcall for reviews on specs open:12:51
dviroel#link https://review.opendev.org/q/project:openstack/watcher-specs+status:open12:51
dviroelI have a particular spec that is under review12:51
dviroelTY all for reviewing it12:51
dviroeli would like to get some feedback on further discussions12:52
dviroelnot sure if you all had time to read it 12:52
sean-k-mooneypartly but not in full12:52
dviroelbut I would like to antecipate, if possible, discussions around the design12:52
amoraleji missed last update12:53
sean-k-mooneyi have 1 or two pending comment but i wanted to better understand the relatship betwen pipeline action plans and audits 12:53
jgilaberme too, I need to set aside to review the last few versions12:53
dviroelso if we have time, we can (next week), discuss more, answer questions12:53
dviroelwe can continue also async in the spec, and discuss in the PTG topic 212:53
dviroelthat will happen in 2 weeks12:53
sean-k-mooneyi was expectign the composaton to be pipeline create audits that prodces action plans but i think its differnt then that in the current spec12:53
sean-k-mooneyso i think that si one area tha ti wotn to udner stand the tade fofs better12:54
sean-k-mooneyif i understand correctly you want ationplans ot optionly be owned by a pipeline is that correct?12:54
sean-k-mooneynote i could be very wrong on that but that a qustion i had after my last skim pass12:55
dviroeltoday action plans are linked to audits only, 1:112:55
sean-k-mooneyyes12:55
dviroelthe spec propose that this could now be linked to a audit pipeline, 1:1 too12:56
sean-k-mooneyare you propsoing changing that as i saw you mention adding the pipeline id ot i think the action plan table12:56
amoralejthat was my understanding12:56
dviroelbut that break additional parameters in the action plan: like strategy_id and goal_id12:56
dviroelwhich are there, not sure why, it is duplicating the info from the audit12:57
sean-k-mooneyyou can have audits without audit templates12:57
sean-k-mooneyso you need to be able to speicy the goal/stragy on the audit dreictly if i recall correctly12:57
amoralejit's a good point about goal_id and strategy_id12:57
sean-k-mooneyrather then alwasy rely on it being in the template12:57
dviroelsean-k-mooney: note that I mentioned Action Plan, it has goal_id and strategy_id12:58
dviroelallong with audit_id12:58
sean-k-mooneyright but action plans are jsut the oputput fo an audit12:58
sean-k-mooneyso those get populated from the audit12:58
sean-k-mooneyi belive that just for trackign but it coudl be looked up tnsitivly via the db relastionship so im not sure it required12:59
dviroelthat one point yes12:59
sean-k-mooneydviroel: anyway the db model was the priamry thing i wanted to look at firt12:59
sean-k-mooneyocne we have the right datamodle i think that will help reason about the rest13:00
dviroelsean-k-mooney: ack, feel free to add comments, and in case that gets more complicated to discuss async, we can have a meeting 13:00
* dviroel time check13:00
sean-k-mooneyyep ill try to find time to do that early next week13:00
dviroelsean-k-mooney: thanks13:00
dviroeland there are 2 other specs that are open for review two13:00
dviroelI already started looking at these too13:01
dviroelnote that we also have some blueprints to be approved13:01
dviroel#link https://blueprints.launchpad.net/watcher13:01
dviroelwhich we can cover in the following meetings13:01
dviroeltoday we don't have more time13:01
chandankumarI have addressed all of the comments on https://review.opendev.org/c/openstack/watcher-specs/+/970220: Add spec for improving watcher-dashboard testing , 13:02
dviroel#topic Volunteers to chair next meeting13:02
dviroelchandankumar++13:02
chandankumarfeel free to take a revisit, thank you!13:02
dviroelI have a question:  do we want to continue with meeting rotation for watcher?13:02
* sean-k-mooney reaslise i never pushed one fo the spec i wrote last year. so i dont need to say i wont have time to work on it this cycle13:02
dviroelI can chair the next one too13:03
sean-k-mooneyare you volenterring to chair it in genral going forward?13:03
sean-k-mooneyalso i assume we will skip this the ptg week right?13:03
dviroelyeah, if nobody wants to chair, i can keep chairing them13:04
sean-k-mooneygiven we ar eover time lets dicuss it next week and assume you will chair that meeting13:04
dviroelcorrect, i am going to cancel the meeting in the PTG week13:05
dviroelyes13:05
dviroellet's wrap up for today13:05
dviroelwe can discuss that afterwards13:05
dviroelthank you all for participating13:05
dviroel#endmeeting13:05
opendevmeetMeeting ended Thu Apr  9 13:05:41 2026 UTC.  Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4)13:05
opendevmeetMinutes:        https://meetings.opendev.org/meetings/watcher/2026/watcher.2026-04-09-12.00.html13:05
opendevmeetMinutes (text): https://meetings.opendev.org/meetings/watcher/2026/watcher.2026-04-09-12.00.txt13:05
opendevmeetLog:            https://meetings.opendev.org/meetings/watcher/2026/watcher.2026-04-09-12.00.log.html13:05
morenodthanks dviroel!13:05
opendevreviewAlfredo Moralejo proposed openstack/watcher master: DNM just check how tests fails when model is modified  https://review.opendev.org/c/openstack/watcher/+/98388415:01
-opendevstatus- NOTICE: Anubis is now deployed on our Gitea backends, and things are back to working normally though you may notice an Anubis screen flash briefly when starting to browse opendev.org; any jobs which failed prior to 15:00 UTC today can be safely rechecked15:35
amoralejsean-k-mooney, dviroel https://zuul.opendev.org/t/openstack/build/a69dd8e153b74841af8ac44c36608c3b from https://review.opendev.org/c/openstack/watcher/+/983884 test fail when adding a new field15:38
sean-k-mooneyamoralej: ah the job has not repoted back yet15:50
amoralejno, i checked in zuul15:51
sean-k-mooneybut that good watcher.tests.unit.api.v1.test_data_model.TestListDataModelResponse failed15:51
sean-k-mooneywhich is the api sample test15:51
amoralejhttps://zuul.openstack.org/status?change=983884%2C115:51
amoralejyep15:51
amoralejlooks good15:51
sean-k-mooneytesttools.matchers._impl.MismatchError: !=:15:52
sean-k-mooneyItems in actual but not in expected:15:52
sean-k-mooney['server_test']15:52
sean-k-mooneyso that fun we prefi them with server15:53
sean-k-mooney 'server_test', form adding test15:53
sean-k-mooneyamoralej: so if we wanted to actully add test we woudl need to update the code to pop it before we return the responce15:54
opendevreviewchandan kumar proposed openstack/watcher-dashboard master: Add Playwright-based E2E testing framework  https://review.opendev.org/c/openstack/watcher-dashboard/+/97035315:54
amoralejsean-k-mooney, yes, that's the plan15:54
sean-k-mooneyare you ok with the existing coverage or do you think we need to do a review of the coverage for all api respocne15:57
opendevreviewchandan kumar proposed openstack/watcher-dashboard master: Add Playwright integration test for skip action workflow  https://review.opendev.org/c/openstack/watcher-dashboard/+/97659415:59
dviroelamoralej: thanks, good that at least would require api code change to have tests passing. 16:26
amoralejsean-k-mooney, i need to check16:29
amoralejbut if needed we can add it together with the filtering code16:29
sean-k-mooneyso at some point we whuld be implementin openapi schema validation for request adn respocne16:30
sean-k-mooneywhen we do that atht will add a second layer of testing of this16:30
sean-k-mooneyas any chagne to the api woudl require a schema chagne as well16:31
sean-k-mooneyso we can build defense in depth here over time16:31
opendevreviewsean mooney proposed openstack/watcher master: run hacking via local hook  https://review.opendev.org/c/openstack/watcher/+/98392519:11
opendevreviewsean mooney proposed openstack/watcher master: run hacking via local hook  https://review.opendev.org/c/openstack/watcher/+/98392519:14
opendevreviewsean mooney proposed openstack/watcher master: Convert remote pre-commit hooks to local pip hooks  https://review.opendev.org/c/openstack/watcher/+/98393320:24
opendevreviewsean mooney proposed openstack/watcher master: Add hook smoke-test script  https://review.opendev.org/c/openstack/watcher/+/98394021:05

Generated by irclog2html.py 4.1.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!