15:00:51 <witek> #startmeeting monasca
15:00:52 <openstack> Meeting started Wed Mar 21 15:00:51 2018 UTC and is due to finish in 60 minutes.  The chair is witek. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:53 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:00:55 <openstack> The meeting name has been set to 'monasca'
15:01:12 <witek> hello
15:01:18 <cbellucci> hi
15:01:21 <sc> howdy
15:01:36 <witek> how is weather in Ireland? :)
15:01:43 <sc> dry
15:01:54 <amofakhar> hello
15:02:06 <jgr> Hello
15:02:23 <jgr> Dry weather in Ireland? That's a first :-)
15:02:53 <witek> we have few topics topics today
15:02:55 <sc> dry = it's not raining right now
15:03:07 <witek> feel free to add if you have some
15:03:08 <jgu_> hello
15:03:20 <witek> #topic Vancouver Forum
15:03:45 <witek> the brainstorming phase for Vancouver Forum is open
15:04:08 <witek> the topics for discussion sessions can be proposed
15:04:56 <witek> Forum are 40-mins. session slots with open discussions
15:05:22 <witek> the goal is to get feedback from operators, users, developers
15:05:30 <witek> or agree on future work
15:05:43 <witek> https://wiki.openstack.org/wiki/Forum
15:05:46 <witek> here more info
15:06:06 <witek> deadline for topic submissions is Apr 2
15:06:28 <witek> e.g. there is already etherpad for self-healing SIG
15:06:34 <witek> which could be interesting
15:06:41 <sc> let me see if I understood: a small PTG with users
15:07:00 <sc> witek: we have to join them for sure
15:07:24 <witek> it's held together with the Summit
15:08:20 <witek> yes, it's somehow like PTG, with more focus on feedback I would say
15:08:29 <amofakhar> is there any possibility it changes some rocky priorities?
15:09:00 <jgu_> 40 minutes daily?
15:09:16 <witek> the output could rather influence next release prios
15:09:47 <amofakhar> ok
15:09:50 <witek> jgu_: the sessions are organised to 40-mins. slots
15:10:36 <witek> everyday through the complete Summit
15:11:07 <sc> who is likely to be in Vancouver? If IIRC we don't have any presentation next summit
15:12:01 <witek> we'll have Project Update and Project Onboarding
15:12:49 <witek> waiting for volunteers to help on these
15:12:56 <sc> OK. What I meant is that if we commit to organized the forum we need to have someone always present
15:13:44 <witek> yes, when proposing the topic normally you should moderate the discussion
15:13:51 <sc> I could volunteer but I have to see if I can travel, budget, approvals... all the big company stuff
15:14:57 <witek> yes, please check
15:15:49 <witek> do you know, who else will attend the Vancouver Summit?
15:16:02 <sc> Added to my todo list, I have to wait till Monday as my boss is off this week
15:16:18 <sc> the home boss granted approval
15:16:35 <amofakhar> we are waiting for approvals
15:16:38 <cbellucci> I am also in the approval phase
15:16:53 <dougsz> I think it's unlikely I'll be able to make it.  I should be able to dial in to relevant sessions.
15:17:29 <witek> I think the sessions are not even recorded
15:17:49 <dougsz> ah
15:18:10 <witek> just etherpad notes
15:18:17 <witek> in most cases
15:18:51 <witek> should we move on?
15:19:02 <sc> yes.
15:19:09 <witek> #topic reviews
15:19:22 <witek> sc: your stage
15:19:31 <witek> https://review.openstack.org/#/c/554304/
15:19:58 <sc> I tested this in devstack and systemd works from the agent. I have to check some interaction with notification
15:20:22 <witek> notification?
15:20:34 <sc> from devstack log
15:20:48 <sc> I see something when the notification starts
15:21:04 <witek> ok
15:21:16 <sc> like a check of services, I'm thinking about something wrong in the devstack/plugin
15:21:27 <witek> so, you basically rework monasca_setup
15:21:34 <sc> yes
15:21:53 <witek> and split to separate 3 services
15:22:09 <sc> yes and one target to manage them all
15:22:30 <sc> monasca_agent now is target that spawn and manage the 3 services
15:22:56 <witek> I guess we'll need some update in documentation as well
15:23:13 <sc> good suggestion witek
15:23:17 <witek> can you do it, or should we split the tasks?
15:23:31 <sc> I prefer to add a new task
15:23:38 <witek> fine
15:23:49 <sc> but I'm going to do
15:24:00 <witek> you've been asking about SysV support
15:24:04 <sc> the big reason this is marked as WIP is that in monasca_setup we have code to manager sysv systems
15:24:10 <sc> could we get rid of this
15:24:10 <witek> :)
15:24:25 <sc> ?
15:24:52 <witek> I think originally monasca_setup supported only SysV
15:24:58 <sc> I mean this is unused code and I don't like leaving code around
15:25:10 <witek> then systemd was added
15:25:29 <witek> does anyone still need support for SysV?
15:25:31 <sc> do you know of any distro supported by OpenStack that is not running systemd
15:25:48 <witek> no
15:26:42 <witek> any opinions?
15:27:45 <witek> I think we can remove this code
15:28:08 <sc> good, I'm doing ASAP and unlock the review
15:28:14 <sc> next topic...
15:28:18 <witek> great, thanks
15:28:27 <witek> https://review.openstack.org/#/c/546956/
15:28:41 <witek> amofakhar: is asking for reviews
15:28:45 <amofakhar> oslo.policy in monasca-api
15:29:00 <amofakhar> yes
15:29:14 <witek> I'll have a look myself in next days
15:29:22 <amofakhar> good thanks
15:29:27 <witek> but would be nice to have someone else as well
15:29:45 <amofakhar> I think sc added himself last week
15:30:31 <witek> amofakhar: anything else on this?
15:30:38 <amofakhar> no thats all
15:30:41 <sc> I took a look then I got interupted by something. I'll complete tonight
15:30:51 <amofakhar> thanks
15:31:16 <sc> amofakhar: ping me on monasca channel if tomorrow you don't have any news from myself
15:31:17 <witek> sc: night is for something else than looking at code ;)
15:31:34 <sc> relaxing before going to sleep
15:31:51 <witek> #topic Rocky prios
15:31:59 <cbellucci> your position is getting worse
15:32:29 <witek> the prio page has been merged
15:32:35 <witek> http://specs.openstack.org/openstack/monasca-specs/priorities/rocky-priorities.html
15:32:57 <witek> it would be great if owners could add short descriptions
15:33:32 <witek> could be one sentence with reference to story or spec
15:34:08 <witek> I have also looked at how we can track progress on prios with StoryBoard
15:34:31 <witek> and haven't found a reasonable way to make it automatically
15:34:57 <witek> I don't want to track all the tasks from StoryBoard, just the prios
15:35:17 <witek> so some kind of tagging would be needed to filter the tasks
15:35:28 <witek> but tags can only be added to stories
15:35:45 <witek> which are not granular enough
15:36:26 <sc> I cannot help
15:36:33 <witek> so I would like to continue to use the board in current shape
15:36:40 <witek> https://storyboard.openstack.org/#!/board/60
15:37:03 <witek> and drag the tasks at least every Wednesday
15:37:34 <witek> that way everyone can see which important changes need review
15:37:48 <witek> or which tasks could be picked up to work on
15:38:10 <witek> also if anyone would like to help updating this board, please ping me
15:38:19 <witek> and I will add the permissions
15:38:36 <witek> any thoughts?
15:39:41 <witek> ok, next topic then
15:40:06 <witek> #topic rejecting future and past measurements
15:40:27 <witek> Tim has provided some more information from Craig
15:40:34 <witek> https://storyboard.openstack.org/#!/story/2001535
15:41:07 <witek> thresholding engine is just dropping these measurements
15:41:21 <witek> and the alarm will stay in UNDETERMINED state
15:41:58 <nseyvet_> that seems OK.
15:42:28 <nseyvet_> Basically, it requires an alarm on the fact that no metrics are coming from a specific agent.
15:42:53 <nseyvet_> so, there is no reason to discard the metrics at the API level
15:43:18 <witek> but the metric would be coming
15:44:32 <witek> oh, so you mean monitoring thresholding engine, right
15:44:33 <witek> ?
15:44:44 <nseyvet_> yes, but it it would be rejected by the threshold engine during the computation
15:44:50 <nseyvet_> yes
15:46:04 <nseyvet_> is it possible to define an alarm in the threshold like "SUM(ALL metrics) == 0 By agent"
15:46:08 <nseyvet_> ?
15:46:32 <witek> don't know, would have to check
15:46:51 <nseyvet_> I guess you could pick any metric really, as basically as soon as that metric does not come in anymore, the total couont of events related to that metric would become 0
15:48:03 <nseyvet_> My guess is that the agent is not a dimension?
15:48:15 <witek> hostname
15:48:20 <witek> in most cases
15:48:32 <nseyvet_> ok, that should work I suppose
15:48:46 <nseyvet_> or?
15:49:27 <witek> I'll think of it, could you also investigate please?
15:50:31 <nseyvet_> yes, I could try to setup a use case
15:50:50 <witek> thanks
15:51:09 <nseyvet_> np
15:51:13 <witek> 10 mins. left
15:51:24 <witek> do we have some other topics?
15:51:56 <sc> https://review.openstack.org/#/c/554918/1
15:52:12 <sc> I don't understand what is this about?
15:53:15 <witek> pypy is supposed to be faster than Python
15:53:43 <sc> it usually is
15:54:03 <sc> so why disabling it?
15:55:08 <witek> as far as I understand, he wants to add this
15:56:15 <sc> the patchset I see it's the other way around, but I may be completelly wrong
15:56:43 <witek> oh, I'm sorry, you're right
15:57:06 <amofakhar> yes it is removed
15:59:02 <sc> I'll ping this guy
15:59:06 <witek> ok
15:59:13 <sc> we have to close
15:59:16 <witek> thank you all for today
15:59:17 <sc> see you
15:59:22 <witek> see you next week
15:59:27 <amofakhar> see you
15:59:27 <witek> bye
15:59:39 <witek> #endmeeting