13:00:37 <witek> #startmeeting monasca
13:00:38 <openstack> Meeting started Tue Mar 24 13:00:37 2020 UTC and is due to finish in 60 minutes.  The chair is witek. Information about MeetBot at http://wiki.debian.org/MeetBot.
13:00:39 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
13:00:41 <openstack> The meeting name has been set to 'monasca'
13:00:51 <chaconpiza> Hi
13:00:55 <witek> hi chaconpiza
13:01:07 <witek> we've got few items on the agenda today
13:01:11 <witek> https://etherpad.openstack.org/p/monasca-team-meeting-agenda
13:01:41 <witek> #topic updating alarm expressions
13:01:56 <witek> bandorf: stage is yours
13:02:04 <bandorf> I added the current status of investigation in the agenda.
13:02:33 <bandorf> I could not yet execute the tests, as discussed 3 weeks ago, due to other priorities.
13:02:57 <bandorf> Bt we've found some issue when further investigating for the issue.
13:03:36 <bandorf> And I still plan to execute these tests (do alarms work properly if we would apply updates for time/times/function?=
13:04:03 <bandorf> Are there any further questions, regarding the descriptions I added as agenda?
13:04:37 <witek> about the undetermined state
13:04:50 <bandorf> Yes?
13:05:00 <witek> although it could indeed indicate some problem in the thresholding engine
13:05:28 <witek> in the real life I'd rather rely on alarms triggered to ALARM state
13:06:50 <witek> also, the described "server down" scenario should probably not be monitored with the `times` operator
13:07:01 <bandorf> So, you think about e.g. an alarm with expression "count(...) == 0" ?
13:07:51 <bandorf> I agree, an explicit server monitoring shouldn't be done that way. That has impact only if there's no explicit alarm for that purpose
13:08:12 <witek> I'd evaluate `host_alive_status` or `http_status` metrics
13:08:19 <witek> or similar
13:08:45 <bandorf> OK, but the same would apply e.g. for services.
13:09:19 <witek> the `times` operator is mainly for avoiding that a short single spike (e.g. in cpu or memory) triggers the alarm
13:09:52 <bandorf> So, you wouldn't even consider this a bug?
13:10:58 <witek> I think it is a bug, and could indicate other problems, but pragmatically could be avoided by careful configuration
13:11:39 <bandorf> OK, agreed. If alarming has been set up 100% perfect, then, there wouldn't be a major impact.
13:12:03 <bandorf> So, at least, we can state that there always should be a work-around
13:12:36 <witek> or even more efficient way of monitoring than described case
13:12:55 <chaconpiza> I recommended our customer to use http_status for monitoring keystone, instead of count + process.
13:14:24 <witek> I could look up what we deploy per default in SOC
13:15:45 <bandorf> OK. To summarize: It is a bug. But it's not a major one. Proper setup of alarm definitions/alarms will check status of servers and services more efficently.
13:16:19 <witek> I'd assume so, but don't know the exact set-up
13:16:46 <witek> the issue with late measurements is interesting
13:17:49 <bandorf> Yes - one of the topics we want to investigate further. It occurs regularly, but not too frequently
13:18:21 <bandorf> As of now, I don't have an explanation. I just know that it occurs independant on load.
13:19:16 <witek> thanks bandorf for the update, good insights
13:19:31 <witek> do we want to move on?
13:19:46 <chaconpiza> One more question related
13:20:18 <bandorf> Go ahead
13:20:19 <chaconpiza> Yesterday bandorf ask me whether there are UI tests like Selenium for monasca-ui
13:20:59 <chaconpiza> I know we don't have them, there was a blocker or a difficulty to implement it?
13:21:21 <chaconpiza> I mean, do other projects have with horizon plugin have them?
13:22:13 <witek> from what I recall, we haven't had IDs for UI elements which are required for testing
13:22:27 <witek> adriancz_ could know more...
13:22:43 <chaconpiza> A-ha...
13:22:57 <witek> we could also reach out to Horizon team and ask about tooling which is avaialble
13:23:11 <witek> and what other projects are doing
13:23:14 <chaconpiza> So far I found this wiki https://wiki.openstack.org/wiki/Horizon/Testing/UI
13:23:21 <witek> s/projects/plugins
13:24:20 <witek> last edit of the page April 2015
13:24:29 <chaconpiza> :S
13:25:05 <witek> but blueprint is marked as completed
13:27:29 <witek> what do you think about reaching out to Horizon team?
13:28:11 <chaconpiza> it is a good idea, I will contact them
13:28:24 <witek> thanks chaconpiza
13:28:27 <bandorf> ok, great
13:28:52 <witek> #topic OpenDev + PTG
13:29:45 <witek> last week OpenStack Committee has officially canceled the Vancouver event
13:30:05 <witek> and want to organize it in a virtual form
13:30:13 <witek> http://lists.openstack.org/pipermail/foundation/2020-March/002854.html
13:31:49 <witek> there is not much planned yet, but they're asking for people who'd like to help
13:32:06 <witek> here the volunteers list
13:32:11 <witek> https://etherpad.openstack.org/p/Virtual_PTG_Planning
13:33:06 <witek> I think it's a good idea to hold Monasca session(s) during that virtual PTG as well
13:34:10 <witek> any comments on that?
13:34:29 <chaconpiza> Virtual sessions will be held through webex ?
13:34:56 <chaconpiza> or each team should define and provide the infra for the sessions?
13:35:37 <witek> I don't know any details, till now there was nothing provided by the organizer but this time the situation is different
13:35:56 <witek> I assume they might provide some infrastructure
13:35:56 <chaconpiza> exactly
13:36:18 <witek> to make it uniform between all the sessions
13:37:44 <witek> #topic PTL nominations season
13:38:17 <witek> today the PTL self-nomination week has started
13:39:00 <witek> I'm not candidating this time
13:39:52 <chaconpiza> I commented internally in Fj that I am interested on that. I will follow the procedure to apply.
13:40:10 <witek> great, thanks for stepping up!
13:40:26 <bandorf> Great!
13:40:52 <chaconpiza> Since next gathering is virtual and the next one is in Berlin, I feel better.
13:41:19 <witek> I believe the change is good for the project and you'll bring new energy
13:42:11 <chaconpiza> From the email you mentioned above: Open Infrastructure Summit in Berlin October 19-23
13:42:53 <witek> right, the date is set, I hope they won't have to cancel this one :/
13:43:44 <witek> I already got an unofficial approval to attend
13:44:31 <chaconpiza> good news
13:45:23 <witek> chaconpiza: please ping me if you have any questions
13:45:37 <chaconpiza> Sure, thanks
13:46:02 <witek> thank you as well :)
13:46:12 <witek> alright, do we have anything else?
13:46:47 <witek> thanks for joining then
13:46:53 <witek> and see you next week
13:46:56 <witek> bye
13:47:02 <adriancz_> thanks
13:47:09 <witek> #endmeeting