12:01:50 <dviroel> #startmeeting watcher 12:01:50 <opendevmeet> Meeting started Thu Nov 6 12:01:50 2025 UTC and is due to finish in 60 minutes. The chair is dviroel. Information about MeetBot at http://wiki.debian.org/MeetBot. 12:01:50 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 12:01:50 <opendevmeet> The meeting name has been set to 'watcher' 12:02:16 <dviroel> hi all o/ 12:02:24 <jgilaber> o/ 12:02:36 <morenod> o/ 12:02:39 <sean-k-mooney> o/ 12:03:15 <dviroel> courtesy ping: amoralej chandankumar rlandy 12:03:22 <rlandy> o/ 12:03:22 <chandankumar> o/ 12:03:23 <amoralej> o/ 12:03:37 <dviroel> oh, the ping works :) 12:03:49 <dviroel> thank you all for joining :) 12:03:53 <dviroel> let's start with today's meeting agenda 12:04:10 <dviroel> #link https://etherpad.opendev.org/p/openstack-watcher-irc-meeting#L27 (Meeting agenda) 12:04:32 <dviroel> we have a couple of topics to cover today 12:04:39 <dviroel> feel free to add your own topics to the agenda 12:04:46 <dviroel> lets start 12:05:00 <dviroel> #topic Announcements 12:05:15 <dviroel> first one is about last week PTG 12:05:49 <dviroel> we had a full week of discussion around different topics and across multiple projects 12:06:10 <dviroel> in watcher sessions we covered: 12:06:45 <dviroel> tech depts, future of integrations, known bugs/issues, improvements to the project, improvements to our testing and new features proposals, among others 12:07:23 <dviroel> the link to the etherpad is 12:07:36 <dviroel> #link https://etherpad.opendev.org/p/watcher-2026.1-ptg 12:07:52 <dviroel> you can also take a look on the summary instead, that I recently sent to the ML 12:08:02 <dviroel> #link https://lists.openstack.org/archives/list/openstack-discuss@lists.openstack.org/thread/CQDEIZKBW6JF4WTE4U5JCIVDNA7FKD7B/ 12:08:27 <dviroel> if you look at the ptg etherpad, on line ~#57 12:08:49 <dviroel> you will find a compilation of action items that was built based on all topics discussed during the week 12:09:12 <dviroel> if you want to help us on any of these topics, please add your name on it 12:09:36 <dviroel> if the item has already a person assigned, you may want to reach that person and see how you can help with that effort 12:10:23 <dviroel> it is very likely that we will bring new discussions to this weekly meeting, as a follow up from our PTG sessions 12:10:42 <sean-k-mooney> i have a second summary here https://gist.github.com/SeanMooney/8a5e8bfc3538917804dfff819c69de10 as well 12:11:02 <sean-k-mooney> i proably wont do a block on the ptg sessions this release like i did last year 12:11:32 <dviroel> sean-k-mooney: thanks for sharing :) 12:11:47 <sean-k-mooney> it takes quite a lot of enery to do that well, dviroel thanks for posting the summery ot the list 12:12:05 <dviroel> ++ 12:12:12 <dviroel> anyone wants to highlight something about the PTG? 12:12:31 <amoralej> thanks for working on the summary dviroel 12:12:44 <sean-k-mooney> i had one tought on reflection after the event 12:13:15 <sean-k-mooney> we choose 3 hrous over 3 days, that left things quite compressed 12:13:30 <sean-k-mooney> teh overlap with nova was also not ideal 12:14:00 <sean-k-mooney> i wonder if we should cosnier withe more shorter sesssions or starting earlier (on the monday) next time 12:14:03 <dviroel> right, most of the team were kinf of using the same timeslots 12:14:15 <dviroel> sean-k-mooney: yes we can 12:14:40 <dviroel> ptg on monday was quiet in the end 12:15:00 <sean-k-mooney> it was a public holiday in ireland so whiel i coudl have attennded i chose not too 12:15:07 <sean-k-mooney> if there were wathcer session i would have obvioulsy 12:15:33 <sean-k-mooney> but i dont know how many other contries had a simialr holiday and if team avoided it as a result 12:15:38 <dviroel> so yeah, next ptg we can try that 12:15:52 <amoralej> i'd vote for doing more sessions, not shorter sessions, i think the conversations were good and i think it's good to give time 12:17:15 <dviroel> we may include one more day in the end 12:18:01 <dviroel> we may avoid conflicts starting earlier and reserving less time on each day 12:18:11 <dviroel> ok 12:18:43 <dviroel> we will have that discussion again next time, when we start booking them? 12:19:05 <dviroel> but that s an important feedback 12:19:12 <dviroel> anything else folks? 12:19:31 <dviroel> we may want to move forward, we have more topics 12:19:49 <dviroel> next on the announcement, a small one 12:19:52 <sean-k-mooney> yep we can move on 12:19:58 <dviroel> a just created a new status etherpad 12:20:06 <dviroel> still working on adding links on it 12:20:13 <dviroel> #link https://etherpad.opendev.org/p/watcher-gazpacho-status 12:20:36 * dviroel created this 5 min before the meeting starts 12:20:46 <dviroel> so we can track our reviews there 12:20:56 <dviroel> any improvement to that etherpad is welcome too 12:21:10 <dviroel> we may want to track backports there too 12:21:12 <dviroel> lets see 12:21:19 <sean-k-mooney> am can i make a request 12:21:23 <dviroel> sure 12:21:35 <sean-k-mooney> can we sue 2026.1 instead 12:21:53 <dviroel> in the etherpad name? 12:21:55 <sean-k-mooney> that technically the offal release name and gazpacho is just the code name 12:21:58 <sean-k-mooney> yes 12:22:01 <dviroel> ack 12:22:07 <dviroel> correct 12:22:08 <sean-k-mooney> ite much eaiser to fined the related ones if we just use the number 12:22:28 <dviroel> since Antelope, the numbering are the official name 12:22:34 <sean-k-mooney> its why i use them for the ptg ethere pads as i often go back years later and refence them 12:22:48 <sean-k-mooney> im not sure we will do that of the status ones but finablity is high on my list 12:22:49 <dviroel> i did had to search how to properly write gazpacho for instance 12:23:00 <dviroel> ack, agree 12:23:11 <dviroel> I can fix that after the meeting 12:23:15 <sean-k-mooney> yes also that i cant spel gazpacho consitently 12:23:21 <sean-k-mooney> +1 12:23:30 <dviroel> thanks for the feedback 12:23:32 <dviroel> ok 12:23:37 <dviroel> moving to the next 12:23:47 <dviroel> #topic Unmaintained branch cleanup 12:23:52 <dviroel> hey jgilaber o/ 12:24:11 <dviroel> #link https://lists.openstack.org/archives/list/openstack-discuss@lists.openstack.org/thread/YK4FRR6LBKZNS3PXFSYH3P3P6HQL4HCS/ 12:24:43 <dviroel> jgilaber recently sent this mail to openstack-discuss ml 12:24:45 <jgilaber> we were talking about umnaintained branches last week with chandankumar and sean-k-mooney and realized we still have the old 2024.1 branch 12:25:03 <jgilaber> so I sent the email to see if anyone was using that branch or had any need of it 12:25:17 <jgilaber> otherwise we can remove it since it's been 2 years since the last patch 12:25:32 <sean-k-mooney> yep 12:25:48 <sean-k-mooney> the default is to remove unless a unmtained branch liason request it 12:26:00 <jgilaber> I wanted to bring it here in case anyone had any objection, if not I'll propose a patch to remove it 12:26:10 <sean-k-mooney> in this case no one has being doing maintainces (it does not have the security bug fix for example) 12:26:15 <sean-k-mooney> so i think we shoudl proceed 12:26:22 <amoralej> +1 12:26:34 <dviroel> +1 on proceed with the proposal 12:27:18 <jgilaber> ack, I'll propose the patch after the meeting 12:27:25 <dviroel> ack, thanks jgilaber 12:27:47 <dviroel> since there is no objections or concerns, lets move to the next topic 12:28:03 <dviroel> #topic Functional tempest jobs for 2025.1 and 2024.2 broken 12:28:09 <dviroel> which is also from jgilaber 12:28:28 <jgilaber> yep, yesterday I submitted a patch to drop the tempest functional job for 2024.1 12:28:41 <jgilaber> #link https://review.opendev.org/c/openstack/watcher-tempest-plugin/+/966146 12:28:53 <jgilaber> and noticed that some functional jobs were consistently failing 12:29:07 <jgilaber> I've created a bug for it https://bugs.launchpad.net/watcher-tempest-plugin/+bug/2130783 12:29:11 <jgilaber> #link https://bugs.launchpad.net/watcher-tempest-plugin/+bug/2130783 12:29:37 <jgilaber> and submitted a patch to make them non-voting while I work on a fix to unblock other patches 12:29:40 <jgilaber> #link https://review.opendev.org/c/openstack/watcher-tempest-plugin/+/966256 12:30:16 <jgilaber> the tldr is that it seems that this commit https://github.com/openstack/tempest/commit/f7470781222524a6a65848721e7f64c6dd5cb8aa is making tox recreate the tempest venv 12:30:35 <jgilaber> and that does not install the watcher-tempest-plugin so it does not find the tests 12:31:37 <sean-k-mooney> ya 12:31:48 <sean-k-mooney> so thise feels like a cobination of things 12:32:01 <sean-k-mooney> newere version fo tox effectivly hash the venenv 12:32:19 <sean-k-mooney> so if it sees that we installed a package into it it will recreat it when we execute a command 12:32:50 <sean-k-mooney> there may be an interactoin with settup toosl as well but this feels like a tempest/devstack bug that we should report to the qa team 12:33:04 <sean-k-mooney> and then fix it ether in our plugin or tempest or devstack so that we do this porperly 12:33:39 <sean-k-mooney> i know there was dicsssion in devstack about how we are currently creating the venv 12:34:38 <sean-k-mooney> lets pick this up with gmaan whne they are online later 12:34:49 <sean-k-mooney> and bring it to the #openstack-qa channel 12:34:58 <sean-k-mooney> we are likely not the only team impacted 12:35:24 <sean-k-mooney> i need to step away for a few minutes so conitnue without me 12:35:29 <dviroel> ack, do we agree on making 2025.1 and 2024.2 non-voting for now? or wait for a fix? 12:35:52 <sean-k-mooney> maybe also 2025.2 i saw that fail on one patch 12:35:57 <sean-k-mooney> but that is my propsoal yes 12:35:59 <chandankumar> +1 for non-voting 12:36:02 <dviroel> i think that it will depend on how quickly this fix will land? 12:36:06 <sean-k-mooney> we just need to be carful with what we merge 12:36:11 <dviroel> yep 12:36:15 <chandankumar> it is failing on my watcher tempest plugin patches also 12:36:48 <dviroel> asking based on how urgent these patches are to land today on in 1 or 2 days 12:37:04 <amoralej> is it affecting jobs only the stable tests in watcher-tempest-plugin or also in the stable branches of watcher ? 12:37:04 <jgilaber> dviroel, maybe it's something simple, but I don't know enough about devstack to guess 12:37:10 <dviroel> ack 12:37:27 <dviroel> lets continue this conversation in the #openstack-qa channel 12:37:38 <dviroel> and we can defer here in the channel later today 12:38:20 <dviroel> thanks for working on this issue jgilaber 12:38:31 <dviroel> and reporting it 12:38:40 <jgilaber> amoralej, I would expect to also affect the watcher stables branches, but I don't think we've had anything recent running there 12:38:50 <amoralej> ack 12:39:01 <dviroel> yeah, ack 12:39:22 <dviroel> ok, so lets move on and continue to track this after the meeting, ok? 12:39:28 <jgilaber> +1 12:39:33 <dviroel> #topic Delete/bulk delete operation on audits/actionplan 12:39:52 <chandankumar> let me take it from here 12:39:59 <dviroel> hey chandankumar this is a follow up from dashboard session from the ptg 12:40:05 <dviroel> :) 12:40:08 <chandankumar> yes correct 12:40:11 <chandankumar> Currently we have openstack optimize audit delete aud1 aud2 aud3 or openstack optimize actionplan delete ap1 ap2 ap3. 12:40:20 <chandankumar> In both case, openstack cli sends a single delete request multiple times to delete the audit or actionplan. 12:40:31 <chandankumar> From codewise, It performs a soft deletion on audit/actionplan, then we need to manually run watcher db purge to delete the audit/actionplan permanently from the DB. 12:40:47 <chandankumar> I saw, When we delete an audit, it does not delete actionplan linked with the audit. Is it expected? or do we we want to extend to perform 12:40:47 <chandankumar> a soft delete on the action plan also? 12:41:36 <dviroel> I would not expect to delete all together 12:41:49 <amoralej> yeah, i'd keep current behavior 12:41:54 <dviroel> the action plan could still exist I think 12:42:05 <dviroel> if we think on a future rollback mechanism 12:42:21 <dviroel> user could use that action plan to rollback something 12:42:33 <dviroel> note: this not exist today 12:42:53 <dviroel> and would not be associated with the audit itself in this case 12:43:00 <amoralej> even for audit purposes, one may want to keep the actionplans visible 12:43:01 <dviroel> only with the action plan 12:43:14 <dviroel> amoralej: yes 12:44:02 <chandankumar> dviroel: amoralej thank you for clarifying it, it make sense to keep it as it is. 12:44:07 <chandankumar> Moving to bulk delete topic 12:44:10 <dviroel> is there a link from action plan to the audit? 12:44:47 <dviroel> in the dashboard? that may break? 12:45:00 <chandankumar> https://paste.openstack.org/raw/bh4xt1eZG8F4GgF2Amo6/ - this is what I have in the cli 12:45:16 <chandankumar> once we delete an audit, in the actionplan list it is set to None 12:45:21 <amoralej> so Audit is set as None 12:45:54 <amoralej> it'd be interesting to check if it's removed in the db or managed in the api 12:46:24 <chandankumar> https://github.com/openstack/watcher/blob/b5725d6ea60d3b7fb2d2b808b261ccdc547df7c4/watcher/api/controllers/v1/audit.py#L741 12:46:42 <chandankumar> based on the code, it perfrom soft_delete, I assume it just update the status to DELETED 12:48:24 <dviroel> when it is soft delete, I guess that it will not appear anymore 12:48:34 <dviroel> but it will be on the db 12:48:40 <amoralej> yes, it is 12:49:04 <amoralej> my question was if the audit field for that actionplan was removed in the db, sorry 12:49:08 <amoralej> in soft_delete 12:49:14 <amoralej> given that cli shows None 12:49:25 <dviroel> ah ok 12:49:37 <amoralej> or it's still in db, but the api is filtering it as it is soft_deleted 12:49:47 <amoralej> just curiosity, np 12:49:47 <chandankumar> amoralej: I will check that and get back on this. 12:50:00 <sean-k-mooney> back 12:50:47 <sean-k-mooney> i woudl expect it to still be in the db 12:50:49 <dviroel> this should be handled by the db 12:50:54 <sean-k-mooney> but it depend on hwo it was hooked up 12:52:22 <dviroel> but the tl;dr; here is to not delete the Action Plan when the audit is deleted, I think that we agree with the current implementation 12:52:34 <chandankumar> yes correct! 12:52:57 <dviroel> chandankumar: you were about to bring another point in this topic 12:53:01 <dviroel> ? 12:53:04 <chandankumar> yes coming to that 12:53:11 <chandankumar> Coming on bulk delete topic, Currently I found reference of bulk delete in swift API and in rest of the project there is no reference of that. 12:53:24 <chandankumar> reference from swift https://github.com/openstack/swift/blob/master/swift/common/middleware/bulk.py 12:53:36 <chandankumar> Since we have a requirement to do bulk archieve for audit/actionplan. How do we want to proceed with implementation? 12:53:45 <sean-k-mooney> ya so its not a common operation that need api support 12:54:00 <sean-k-mooney> well it depends 12:54:05 <amoralej> it would be asynchronous task ? 12:54:06 <chandankumar> If an user passes openstack optimize audit delete aud1 aud2 aud3 , Does it will call bulk delete api, it will do a single api call to delete all the passed audits or it will delete all the audit and perform soft delete from the db? 12:54:30 <sean-k-mooney> archiving an action plan shoudl archive teh actions assocated wtih the plan 12:54:57 <sean-k-mooney> archiving a one shot audit coudl do the same, but that is less clear that it shoudl 12:55:12 <sean-k-mooney> certenly if we want the behavior to be the same for contious audit we woudl nto want ti too 12:55:35 <amoralej> archiving an actionplan already archives the actions 12:55:40 <sean-k-mooney> but yes at the api level we could take the uuids as a list in the query stirng 12:55:59 <sean-k-mooney> form a http point of view the delete method is not expected ot have a body 12:56:29 <sean-k-mooney> so if we dont want to hit the quary string lenght limit 12:57:05 <sean-k-mooney> the othe roption wouls be a post to a new api endpoint with the auits or action plans to archive listed in json in the boday 12:57:54 <sean-k-mooney> the final optoin i see is takign a `cacade` or simialr query arg to the audit delete 12:58:04 <sean-k-mooney> to opt into archiing the action plans as well 12:58:22 <chandankumar> currently we donot archive actionplan if we archieve audit 12:59:29 <dviroel> we could, if the user decide to do that with an option, but yeah, we don't do by default 12:59:43 <amoralej> i like the idea of making it an option 12:59:44 <dviroel> i mean, we don't support that 12:59:57 <sean-k-mooney> so that 3 related optiosn `DELETE /audit/<uuid>?cacade=true` `DELETE /actionplan/?uuids=....` or `POST /archive with a json body` 13:00:39 <sean-k-mooney> the ohter option woudl be do do it client side only which im not sure is correct 13:01:37 <sean-k-mooney> if we were to just do it in the watcher dashbaord it woudl have to be server side in the plugin not in javascript 13:01:39 * dviroel time check 13:01:57 <sean-k-mooney> i think there is enoguh dicusion here to show that in any case this need a spec 13:02:01 <sean-k-mooney> do we agree? 13:02:09 <amoralej> i'd expect horizon to provide some support for bulk? 13:02:10 <dviroel> yes 13:02:21 <chandankumar> from dashboard point of view , use can click on bulk archieve button to archieve all audit? 13:02:28 <dviroel> yes -> agree that we need a spec 13:02:41 <amoralej> +1 to spec 13:02:48 <sean-k-mooney> im not sure we need this at all by the way 13:03:08 <sean-k-mooney> my general prefence woudl be to have an expiry time on the audit/action plans 13:03:14 <amoralej> i see it as a nice-to-have requirement, tbh 13:03:28 <sean-k-mooney> both for howlong an pending action plan is retains and completed ones 13:03:45 <sean-k-mooney> so a time to live for unappoved action plans 13:03:57 <sean-k-mooney> and an expiry time before they are auto archived 13:04:10 <sean-k-mooney> but we can talk about the usecase in the spec 13:04:16 <dviroel> expiry time is also a good idea from my pov 13:04:17 <sean-k-mooney> and then desing the correct feature or feature to adress that 13:05:04 <dviroel> chandankumar: yeah, we need more time to think about and the spec would be a good starting point, describing the uses cases and possible solutions 13:05:22 <chandankumar> + for expiry time, Does is it going to be a seperate feature? 13:05:33 <dviroel> and maybe we can revisit in future meetings if needed 13:05:47 <chandankumar> I got more info for the spec now, I will start working on that 13:06:05 <sean-k-mooney> we can scope that doen in the spec 13:06:13 <sean-k-mooney> if its too complicate we can split it 13:06:19 <dviroel> +1 13:06:22 <chandankumar> +1 13:06:26 <chandankumar> thank you all ! 13:06:29 <sean-k-mooney> if not we can have one spec for audit/action lifetimes 13:06:34 <dviroel> thanks chandankumar 13:06:38 <sean-k-mooney> and cover the watcher ui and cli imapcts 13:06:46 <dviroel> so we are out of time 13:07:00 <dviroel> there are 2 links in the reviews topic 13:07:18 <dviroel> #topic Reviews 13:07:25 <dviroel> #link https://review.opendev.org/c/openstack/watcher/+/958766: Remove watcher_notifications from default value 13:07:26 <dviroel> and 13:07:33 <dviroel> #link https://review.opendev.org/q/topic:%22bug-2126767%22+and+status:open service_monitor for decision-engine 13:07:43 <dviroel> we don't have time to go into details 13:07:56 <dviroel> but we should be looking at those as requested 13:08:15 <dviroel> amoralej: we can get you a topic at the start of next meeting if needed 13:08:16 <amoralej> second one was just to bring those to your atenttion 13:08:22 <dviroel> amoralej: ack, thanks 13:08:29 <amoralej> we can discuss in the reviews 13:08:37 <dviroel> #topic Volunteers to chair next meeting 13:08:51 <jgilaber> I can do it dviroel 13:08:57 <dviroel> jgilaber: TY 13:09:10 <dviroel> so 13:09:10 <dviroel> let's wrap up for today 13:09:14 <dviroel> we will meet again next week 13:09:23 <dviroel> thank you all for participating 13:09:26 <dviroel> #endmeeting