Thursday, 2026-01-29

opendevreviewchandan kumar proposed openstack/watcher-specs master: Add spec for improving watcher-dashboard testing  https://review.opendev.org/c/openstack/watcher-specs/+/97022006:40
opendevreviewJoan Gilabert proposed openstack/watcher master: Prepare to use openstacksdk instead of novaclient  https://review.opendev.org/c/openstack/watcher/+/97471009:19
opendevreviewJoan Gilabert proposed openstack/watcher master: Complete migration from novaclient to openstacksdk  https://review.opendev.org/c/openstack/watcher/+/97492409:19
opendevreviewJoan Gilabert proposed openstack/watcher master: Remove usage of novaclient from Watcher  https://review.opendev.org/c/openstack/watcher/+/97492509:19
chandankumarIrc meeting starts in 5 mins.11:55
chandankumar#startmeeting watcher12:01
opendevmeetMeeting started Thu Jan 29 12:01:17 2026 UTC and is due to finish in 60 minutes.  The chair is chandankumar. Information about MeetBot at http://wiki.debian.org/MeetBot.12:01
opendevmeetUseful Commands: #action #agreed #help #info #idea #link #topic #startvote.12:01
opendevmeetThe meeting name has been set to 'watcher'12:01
chandankumaro/12:01
dviroelo/12:01
jgilabero/12:01
chandankumarcourtesy ping list: sean-k-mooney rlandy amoralej morenod 12:01
amoralejo/12:01
morenodo/12:01
rlandyo/12:01
chandankumarthank you for joining the meeting.12:02
chandankumarlet's get started with the today's meeting agenda12:02
chandankumar#link https://etherpad.opendev.org/p/openstack-watcher-irc-meeting#L2312:02
chandankumarfeel free to add your topic to the agenda12:02
chandankumar#topic Reset the DPL model for watcher project12:03
chandankumar#link https://review.opendev.org/c/openstack/governance/+/97492612:03
chandankumarSince new openstack release is coming, TC has proposed the above patch to reset the DPL model for watcher project12:04
chandankumarfor 2026.2, Do we want to continue the DPL model or anyone wants to step up for Watcher PTL?12:05
dviroelo/12:05
chandankumarIf we want to continue the DPL model in that case, we need to find liasions for different tasks.12:05
dviroeli do have plans to add my candidacy for PTL for next cycle12:06
dviroelso i would propose to allow the DPL reset patch merge12:06
chandankumardviroel: wonderful!12:06
chandankumardviroel: I think We still need liasions for different task to assist PTL during the whole release.12:07
dviroelright!12:07
dviroeli would still need assistance for liaisons in release and security context at least :) 12:08
chandankumarjgilaber: and me are the current liasons for release and sean-k-mooney was looking after security.12:08
chandankumarwe can check with them if they want to continue for next release.12:09
jgilaberthanks for stepping up dviroel I can continue as release liaison12:09
chandankumarI also happy to help anywhere if you need more hands.12:10
dviroeljgilaber: tks, we can chat about that after elections then, which start in the following weeks12:10
dviroel#link https://governance.openstack.org/election/12:10
dviroelsomeone needs to be elected first :) 12:10
chandankumarPTL nominations will start next week i.e 4th Feb.12:10
dviroel+112:11
jgilaberfor now we just need to +1 the TC patch to reset to PTL and that it right?12:11
dviroelright12:11
jgilaberack thanks12:12
chandankumareveryone feel free to add your votes on https://review.opendev.org/c/openstack/governance/+/97492612:12
dviroelack12:13
chandankumarI think we are all sorted on this topic. we have a upcoming PTL nomination from dviroel and new liasons. Anything more to discuss on this12:13
rlandy+112:13
dviroelwe can continue I think12:13
chandankumarnext topic12:14
chandankumar#topic Deciding watcher-dashboard integration testing framework12:14
chandankumarWe are already discussing about watcher-dashboard integration testing poc in multiple meetings.12:14
chandankumarI have updated the spec with selenium vs playwright comparison for a watcher developer point of view: 12:14
chandankumar#link https://review.opendev.org/c/openstack/watcher-specs/+/970220/10/specs/2026.2/approved/improve-watcher-dashboard-ui-testing.rst#L16312:15
chandankumarhtml version: https://abad4d81dfb03b4f334d-70b013fe9ea05f673313d916271f541a.ssl.cf5.rackcdn.com/openstack/b38087f5bdb4410294bd985e1d93a099/docs/specs/2026.2/approved/improve-watcher-dashboard-ui-testing.html#playwright-vs-selenium-poc-notes-watcher-developer-view12:15
chandankumarI wanted to check with the team to make a final decision on integration testing framework12:15
chandankumarso that Devs can start adding the integration tests for watcher-dashboard features.12:15
morenodI've been playing with playwright. my vote is for it12:16
jgilaberI have no direct experience, but my impression from reviewing the POCs and spec, and from previous discussions it seems playwright has some advantages over selenium12:17
jgilaberso +1 from me12:17
amoralejeverything i saw from playwright seems to sounds as better or equal than selenium in testing capabilities and user experience12:17
chandankumarHow do we want to vote on this? Do we need +1 on the spec itself with the test framework name?12:17
amoralejby user, i mean developers :)12:17
amoralejand also for integration in CI jobs12:18
sean-k-mooneywell normally we dont vote persay12:18
jgilaberimo we should update the spec to propose playwright as the test framework12:18
sean-k-mooneymore lazy concenus12:18
jgilaberand list selenium as a possible alternative12:18
sean-k-mooneybut it seam we are gravatating to playright12:18
dviroelchandankumar: you did a great job on both PoCs +1 - taking over the notes I have the impression that playwright is best 12:18
amoralejand i guess we will be able to handle adding it to the global-requirements 12:19
sean-k-mooneyso what i woudl suggest is to update the spec to propsoe using that and move selenium to the alternives section12:19
chandankumarwonderful, everyone is in favor of selenium.12:19
sean-k-mooneyamoralej: ya that should not be hard but we should start the process early12:19
amoralej+112:19
chandankumar*everyone is in favor of playwright12:19
sean-k-mooneydont you mean playrgiht12:19
sean-k-mooney:)12:19
chandankumarI will update the spec with playwright and move selenium as a alternative.12:20
dviroel+112:20
chandankumarThank you all for taking the time to go over spec and poc and finalizing the tool.12:20
sean-k-mooneyok then once you have updated the spec to say that we can review it and start working on the reqiurements change and reviwign the changes once the master banch is 2026.212:20
chandankumar+112:21
sean-k-mooneyare we finished with this topic12:21
sean-k-mooneyif so before we go to reviews per the adgenda i woudl like to talk abotu some release stuff12:21
amoralejthanks chandankumar++12:22
chandankumaryes12:22
chandankumarsean-k-mooney: sure12:22
chandankumar#topic release12:22
chandankumarsean-k-mooney: please go ahead12:22
sean-k-mooneyok so first release topic. let try and do some stable release next week. my proposal is lets do some stable reviews today and tomorrwo and then work on proposing the release early next week12:23
dviroel+112:23
jgilaber+112:23
chandankumar+112:23
sean-k-mooneysecond related topic is we are startign to get close to FF so there are some release tasks liek starting on the prelude and marketing highlight that we will need to start soon12:23
sean-k-mooneywe have another week or two before we actully need to do those so that just an fyi that we shoudl consult the chronalgical release guide and prepare for those tasks in the next few weeks12:24
sean-k-mooneywe can loop back to that in the next meeting or the one after12:24
sean-k-mooneythat was basically all i wanted to bring up12:25
chandankumarDo we need a voluteer for creating prelude and marketing highlights?12:25
rlandyI can try12:26
jgilaberI can also help with that12:26
sean-k-mooneythey offially are duties of the release liason but anyone can propose them12:26
sean-k-mooneyit just need a release liason to approve the marketing highlight patch in the release repo12:26
jgilaberjust checked, FF is the week of Feb 23 - Feb 2712:26
sean-k-mooneyyep and the non clinet lib freeze if feb 9th12:27
sean-k-mooneywell week of feb 9th12:27
chandankumarrlandy: jgilaber thank you for volunteering. 12:28
sean-k-mooneythe prelude shoudl be doen at or aroudn FF but before RC1 the hilights shoudl be the first week of march at the latest12:28
chandankumarOnce the prelude review is up, we can follow up in next meeting .12:28
chandankumarok12:29
chandankumarjust to summarize: We will focus on reviews for stable branches this week to prepare a new stable release next week. Release liason will work on crafting the prelude and marking stuff.12:31
chandankumarAnything else to discuss on this topic12:31
sean-k-mooneyya so we have 3-4 weeks for those artifact to be created but its just good to start a head of time so we are not in a rush12:32
sean-k-mooneywe can likely move to the reviews i think we are dont with this topic12:32
chandankumar#topic reviews12:32
chandankumar#link     Enable extra zone migration tests on master branch | https://review.opendev.org/c/openstack/watcher/+/97038112:32
jgilaberI added this patch12:34
chandankumarAs the review says it enables extra zone migration tests on master branch, Depends on already got merged.12:34
jgilaberit's a small patch adding some zone migration tests that we put behing a config flag in the tempest plugin12:34
jgilaberthey work in master, but need some bugfixes to be backported in the stable branches12:35
jgilaberthe patch is only for master, I can backport it as well once/if we merge those bugfixes12:35
dviroel+112:35
sean-k-mooneylooks fine to me12:35
sean-k-mooneydouble workflow :)12:36
jgilaberthanks!12:36
dviroellol12:36
chandankumarwonderful12:36
chandankumar#link     stable 2025.2 open reviews: https://review.opendev.org/q/project:openstack/watcher+branch:stable/2025.2+is:open12:36
dviroeli added these earlier12:37
dviroeljust to reminder about stable branches reviews12:37
chandankumarwe have many reviews which needs one more +2 and +w in 2025.212:37
sean-k-mooneyah i see you found time to review them last week cool12:37
dviroelI did some reviews on 2025.2 and will start 2025.1 12:37
chandankumar#link https://review.opendev.org/q/project:openstack/watcher+branch:stable/2025.1+is:open12:37
sean-k-mooneycool i will look at them after the meeting as well12:38
chandankumareveryone please have a look at 2025.1 and 2025.2 series, it will help us to make new release next week12:38
chandankumarthat's it from review12:38
chandankumar#topic Bugs12:39
chandankumarI can see two untriaged bugs in watcher-tempest-plugin12:40
chandankumar#link  Race condition in test_continuous_audit_actionplan_superseding https://bugs.launchpad.net/watcher-tempest-plugin/+bug/2136124  12:40
sean-k-mooneyit sound valid proably medium12:41
sean-k-mooneyany instablity causes ci impact that we want to avoid12:41
chandankumarthanks sean-k-mooney 12:42
chandankumartriaged !12:42
sean-k-mooneymy guess is we need a retry loop to stablise it12:43
amoralejor cancel the old actionplans after the new one is created instead12:43
sean-k-mooneyperhaps 12:44
sean-k-mooneyi guess either would be ok12:44
dviroelamoralej: do you have a patch or plan to work on one?12:44
amoralejnot12:44
amoralejjust reported, tbh12:45
amoralejbut I'd say, the alternative is to move cancelling around https://github.com/openstack/watcher/blob/2ba9caae744db2f62e6a57e25069ac364ec54ffc/watcher/decision_engine/audit/base.py#L9412:45
dviroelack, just want to know to update the assignee 12:46
amoralejalthough, we may have the oposite race condition ...12:46
amoralejhaving two recommended APs for a while12:46
amoralejthe retrying loop doesn't sounds bad now :)12:47
sean-k-mooneywe shouldnt12:47
sean-k-mooneybecuase the previousl recommened plan shoudl be move to superceeded12:48
sean-k-mooneywe may have a bug if we we are not doing both in a single transaction i guess12:48
sean-k-mooneythat woudl be a watcher bug12:48
amoralejyes, that's the issue12:48
sean-k-mooneyas i dont think we are ever meent to have more the 1 recomemend action plan12:49
amoralejno, we shouldn't12:49
amoralejwe need to make both state changes atomically12:50
amoralejadding the new AP and changing the state of the old one12:50
sean-k-mooneyyep hence same transaction or we woudl neeed a db trigger12:50
sean-k-mooneywe dont tend to use triggers in openstack so same transacton woudl be the correct fix12:50
sean-k-mooneylets revisit this when someone starts working on the bug fix but we might end up fixing it in watcher instead12:52
sean-k-mooneywe shoudl dig a little deeper before updating the test12:52
amoralejbtw, we don't use SUPERSEDED state for this use case (old APs in a continuous Audit, when a new one is created) but CANCELLED12:52
amoralejwe use superseded only for time based expiration iirc12:52
sean-k-mooneyso cacled for an action plan si only valid if the action plan is runnng12:53
sean-k-mooneyand if it running it should not be canceld by the creation fo a new action plan12:54
sean-k-mooneyit should run to completion12:54
amoralejhttps://docs.openstack.org/watcher/latest/architecture.html#action-plan-state-machine12:54
amoralejAPs can be moved from RECOMMENDED to CANCELLED12:54
sean-k-mooney if and only if a admin does it12:55
amoralejthat's actually used when an user wants to avoid running it12:55
amoralejyes, exactly12:55
sean-k-mooneyright but if we are testing supperceded as the name impleis the tempst shoudl not call cancel at all12:55
amoralejwell, in current implementation alse for continuous audits when a new AP is created12:55
amoralejah, superseding is actually missleading12:56
amoralejcontinuous actionplans do not use superseded status right now12:56
sean-k-mooneyright so that a bug12:56
sean-k-mooneya diffent bug but a bug12:56
amoralejyeah, it'd may make sense use superseded instead of cancelled for this12:56
amoralejdunno if that's documented tbh12:57
sean-k-mooneywatcher should not move it to canceled when the new action plan is created12:57
sean-k-mooneywell that what the state machine doc says we shoudl be doing12:57
dviroelyeah, that's my understanding to, from the docs/state machine12:58
sean-k-mooneyi dont think there shoudl be an state change form recommend to cancheld ever for what its worth12:58
amoralejso the bug is "Watcher is setting previous APs as CANCELLED instead of SUPERSEDED when a new one is created in a CONTINUOUS audit" ?12:59
sean-k-mooneyyou should be able to go form recommend to deleted directly or recommend to supseeded automaticlly12:59
sean-k-mooneyright cnceled should only be reachable form the ongoing state in my view13:00
sean-k-mooneywell perhaps also form bending13:00
amoralejor by admin api call, right?13:00
sean-k-mooney*pending13:00
sean-k-mooneyno13:00
chandankumarhere is the action plan state machine diagram https://github.com/openstack/watcher/blob/master/doc/source/image_src/plantuml/action_plan_state_machine.txt13:01
sean-k-mooneyim saying that is not somethign we shoudl be supproting13:01
amoralejwe need it with the current wf13:01
amoralejas we can not delete a recommended AP13:01
sean-k-mooneyright im saying that workflow is invlaid semanticly13:01
sean-k-mooneythe replacement workflow would be recommend to deleted directly or recomened eto superseeded13:01
amoralejtbh, i think it's fine a user can CANCEL it before deleting it13:02
amoralejdeleting it means, no way to see it anymore13:02
sean-k-mooneyi dont think cancel is valid if it not running (pending or ongoing) that my general point if it did not start it cant be canceled13:02
dviroelyeah, what looks weird is to Cancel something that was not even running/started13:02
amoralejsomeone may be interested in running an audit, cancel but keep it to see the efficacy indicators, i.e. and compare to a new execution of the audit13:03
amoralejwe may use a different state name, but imo, the use case is valid13:03
sean-k-mooneyright so a user shoudl be able to delete the recommened action plan. and watcher shoudl automaticlly mvoe ti to superseded instead of cancled when a new action plan is created IMO13:03
amoralejanyway, I will report this bug, no need to agree on a solution now13:04
sean-k-mooneylets leave this for now and leave a note in the bug13:04
sean-k-mooneyack13:04
dviroel+113:04
chandankumarI am moving to last topic as we passed the meeting time limit13:05
chandankumar#topic volunteer for next meeting13:05
chandankumaranyone up for chairing the next meeting?13:05
morenodI will13:06
chandankumarthanks morenod 13:06
chandankumar#endmeeting13:07
opendevmeetMeeting ended Thu Jan 29 13:07:04 2026 UTC.  Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4)13:07
opendevmeetMinutes:        https://meetings.opendev.org/meetings/watcher/2026/watcher.2026-01-29-12.01.html13:07
opendevmeetMinutes (text): https://meetings.opendev.org/meetings/watcher/2026/watcher.2026-01-29-12.01.txt13:07
opendevmeetLog:            https://meetings.opendev.org/meetings/watcher/2026/watcher.2026-01-29-12.01.log.html13:07
chandankumarthank you all for joining the meeting!13:07
dviroelthanks chandankumar 13:07
jgilaberthanks chandankumar!13:07
morenodthanks chandankumar++13:07
amoralejthanks chandankumar++13:07
amoralejbtw, the current workflow is documented in https://github.com/openstack/watcher/blob/master/doc/source/user/continuous_type_audit.rst?plain=1#L84-L86 we need to fix it if we change it13:10
sean-k-mooneythat new docs that chandan wrote 4 months ago https://github.com/openstack/watcher/commit/3c8bc6be6236335d6558bd6576f3f72351f224f113:11
sean-k-mooneyso that not nessicarly correct13:11
sean-k-mooneyits documenting the current flow yes but not the intended flwo/api contract13:11
amoralejyep, no problem with changing it13:11
sean-k-mooneywe would have to look at the orginal specs13:12
amoralejreported https://bugs.launchpad.net/watcher-tempest-plugin/+bug/213934413:13
sean-k-mooneyhttps://specs.openstack.org/openstack/watcher-specs/specs/newton/implemented/continuously-optimization.html13:15
sean-k-mooneyTo keep track of the triggered audit, notification has to be pushed on the message bus every time the audit is re-triggered. When a new action plan is proposed, Watcher should cancel all the previously generated action plans (and actions) with same Audit Template become obsolete and therefore their state should be changed to CANCELLED.13:16
amoralejyep13:16
sean-k-mooneyya so that soudn like a design bug13:16
sean-k-mooneybut  if we want to change it it more of a feature or enhacment13:16
sean-k-mooneyso i guess we need to condier the upgrade impact13:17
sean-k-mooneyfor now we may want to rename the test to make this clearer13:17
amoralejIMO, it'd be more correct to use SUPERSEDED but it's low priority enhancement13:18
*** user is now known as Guest78813:18
opendevreviewJoan Gilabert proposed openstack/watcher master: Prepare to use openstacksdk instead of novaclient  https://review.opendev.org/c/openstack/watcher/+/97471018:59
opendevreviewJoan Gilabert proposed openstack/watcher master: Complete migration from novaclient to openstacksdk  https://review.opendev.org/c/openstack/watcher/+/97492418:59
opendevreviewJoan Gilabert proposed openstack/watcher master: Remove usage of novaclient from Watcher  https://review.opendev.org/c/openstack/watcher/+/97492518:59
opendevreviewJoan Gilabert proposed openstack/watcher master: Change id field to uuid when appropiate  https://review.opendev.org/c/openstack/watcher/+/97520018:59
opendevreviewMerged openstack/watcher master: Enable extra zone migration tests on master branch  https://review.opendev.org/c/openstack/watcher/+/97038120:42
opendevreviewMyles Penner proposed openstack/watcher master: Fix Sphinx build failure with Python 3.14  https://review.opendev.org/c/openstack/watcher/+/97523122:34
opendevreviewMerged openstack/watcher stable/2025.2: Ignore cinder notifications if the Storage model is empty  https://review.opendev.org/c/openstack/watcher/+/96323723:18

Generated by irclog2html.py 4.0.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!