15:00:49 <witek> #startmeeting monasca
15:00:49 <openstack> Meeting started Wed Oct 24 15:00:49 2018 UTC and is due to finish in 60 minutes.  The chair is witek. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:50 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:00:52 <openstack> The meeting name has been set to 'monasca'
15:01:02 <joadavis> greetings
15:01:03 <witek> Hello everyone
15:01:10 <witek> hi joadavis
15:01:15 <dougsz> hi all
15:01:19 <truongnh-fujitsu> Hi witek
15:01:31 <witek> hi dougsz, truongnh-fujitsu
15:01:35 <TuanVu> Hi everybody
15:01:43 <witek> let's start
15:01:46 <witek> the agenda:
15:01:51 <witek> https://etherpad.openstack.org/p/monasca-team-meeting-agenda
15:02:08 <witek> #topic delete alarms API
15:02:21 <witek> https://storyboard.openstack.org/#!/story/2004099
15:02:38 <witek> joadavis: you've added this story recently
15:03:09 <joadavis> yes, this was a simple story someone on our team brought up.  They were deleting a node and wanted to clean out any alarms related to the node
15:03:27 <joadavis> sounded reasonable so I added a story
15:04:00 <witek> use case could also be valid after removing vm
15:04:02 <joadavis> it is currently possible to get a list of the alarms then delete one by one
15:04:08 <joadavis> yes
15:04:46 <witek> sounds like a good improvement to me
15:04:56 <joadavis> I'm not working on the story, so anyone is welcome to comment or pick it up
15:05:27 <witek> do you think we would need a short spec for this?
15:06:09 <witek> it's changing the API, but in a minimal way
15:06:41 <joadavis> It seemed simple so I hadn't thought about it, but it may be beneficial to have a spec to give some examples.
15:07:02 <witek> yes, spec would be nice
15:07:22 <witek> but we can also just start with updating the documentation
15:07:32 <witek> and then following with the implementation in the same change
15:08:23 <joadavis> good
15:08:44 <witek> I'll add the story to our board
15:08:53 <witek> https://storyboard.openstack.org/#!/board/111
15:09:03 <witek> so someone can pick it up
15:09:25 <witek> thanks for the idea joadavis
15:10:08 <witek> anything else on this topic?
15:10:16 <joadavis> I'll pass on the thanks to Gloria who brought it up. Nothing else
15:10:29 <witek> :) thanks Gloria
15:10:48 <witek> #topic proactive HA for VMs
15:11:11 <witek> hello Fujitsu team
15:11:17 <TuanVu> Hi everybody,
15:11:17 <TuanVu> I’m Tuan, from Fujitsu Vietnam
15:11:22 <truongnh-fujitsu> Hello witek
15:11:31 <TuanVu> I, truong and annp_ are working on Proactive HA feature
15:11:37 <truongnh-fujitsu> I'm Truong, from Fujitsu Vietnam
15:11:55 <TuanVu> We're just wondering if Monasca team have any plan or spec for supporting Proactive HA for VM?
15:12:38 <witek> we have implemented auto-scaling with Heat
15:12:58 <witek> but I guess, you're interested in more than this
15:12:59 <TuanVu> thank you, witek
15:13:00 <joadavis> (side note - I was working on a fix for the autoscaling with heat in stable/pike this week)
15:13:12 <truongnh-fujitsu> We are investigating the metrics using in Live Migration in Masakari, too
15:13:36 <TuanVu> yes, we would like to focus on Proactive HA
15:13:53 <TuanVu> auto-scaling is different topic, I think
15:14:13 <witek> correct
15:14:23 <witek> we've discussed it during the last PTG
15:14:38 <TuanVu> If Proactive HA is not supported at this moment, can we propose for Monasca a list of metrics which are necessary for Proactive HA?
15:15:18 <witek> sure, I've seen you added an etherpad to collect these
15:15:23 <TuanVu> here's a brief examples for the list of metrics
15:15:24 <TuanVu> https://etherpad.openstack.org/p/Monasca_-_supporting_Proactive_HA_for_VM
15:15:54 <TuanVu> yes, it would be great if we have contribution from our community as well
15:16:28 <TuanVu> Since we’re newbie in Monasca, we really appreciate your help :)
15:16:28 <witek> I tried to put the plugins, which provide these metrics to the etherpad
15:16:28 <TuanVu> Thank you guys in advance!
15:16:45 <TuanVu> that's awesome! Thank you very much, witek
15:17:11 <witek> host status can be checked with host_alive_plugin and/or http_check
15:17:31 <witek> net.int_status is a standard system check
15:17:55 <witek> libvirt metrics are well documented in a separate document
15:18:23 <TuanVu> thank you, I'm checking the document which you noted
15:18:37 <TuanVu> what about CPU temperature?
15:18:38 <witek> we don't have native support for IPMI
15:19:03 <joadavis> and anything you might want from Ceilometer you may find an equivalent metric already in Monasca or you can use Ceilosca to transfer the samples to Monasca
15:19:16 <witek> if you need them, we have few options to address it
15:19:38 <witek> joadavis: yes, thanks for addition
15:19:53 <witek> for IPMI:
15:20:07 <witek> 1) we could add a new plugin to monasca-agent
15:20:52 <witek> implementation should be pretty much straight forward
15:21:28 <witek> 2) you could you Prometheus exporter and collect measurements with our Prometheus plugin
15:22:14 <witek> 3) you could use Ceilometer to collect these and publish them to Monasca using monasca-ceilometer
15:23:36 <witek> the documentation describing adding plugins is available here:
15:23:38 <witek> https://github.com/openstack/monasca-agent/blob/master/docs/Customizations.md
15:23:42 <TuanVu> thanks a lot for the answer with lots of details, witek :)
15:24:10 <TuanVu> we'll research more on these options
15:24:55 <witek> feel free to ping us if you need any help
15:25:13 <witek> btw, will you travel to Berlin?
15:26:13 <TuanVu> thanks, we'll let you know if there's any concern
15:26:26 <TuanVu> no, I will not
15:27:38 <truongnh-fujitsu> In the upcoming summit, we will not go to Berlin
15:28:12 <witek> do you have a description of your use case, which would be suitable for publishing?
15:28:52 <witek> we could create a story in StoryBoard and add tasks for openstack/self-healing and monasca-agent, monasca-notification repos
15:29:34 <truongnh-fujitsu> Now, we are collecting all of the metrics using in Monasca, we will propose our use cases soon
15:29:45 <TuanVu> the detailed description is not ready yet
15:30:11 <TuanVu> thanks for your suggestion, witek
15:30:17 <witek> that's fine, thanks
15:31:08 <witek> thanks for bringing this up, and please keep us updated
15:31:26 <witek> #topic Stein-1 milestone
15:31:55 <witek> tomorrow we reach the first milestone in Stein
15:32:14 <witek> I'll tag the repos if applicable
15:32:38 <witek> we should get our important specs merged soon
15:33:06 <joadavis> yes, I have been neglecting the spec revisions I was working on
15:33:34 <witek> I thought in particular about events-listener and merging APIs
15:34:30 <joadavis> I fixed the autoscaling bug that was eating my time for the last 2 days, so I can make some spec changes today
15:34:45 <witek> I've talked with the team and it seems we'll be able to work on merging APIs
15:35:13 <witek> joadavis: btw, do we have any documentation about heat autoscaling?
15:35:49 <joadavis> it was all new to me before Monday. :)  But there is a sample already documented in the hot docs...
15:36:04 <witek> hot docs? :)
15:36:53 <joadavis> HOT - heat templates. :)
15:36:59 <witek> I think use case in self-healing repo would be very sexy
15:37:03 <joadavis> Rico did a related article... http://superuser.openstack.org/articles/how-to-auto-scale-a-self-healing-cluster-with-heat/
15:37:42 <joadavis> and we just got https://github.com/openstack/heat-templates/blob/master/hot/monasca/autoscaling.yaml to work on pike, after applying a WIP patch
15:38:08 <joadavis> https://wiki.openstack.org/wiki/Heat-Monasca-Auto-Scaling is also relevant
15:38:37 <witek> does Rico's article mention triggering with Monasca?
15:39:20 <joadavis> no, he was using another service
15:40:04 <witek> the wiki page is pretty old, do you know if it's up to date?
15:40:47 <dougsz> witek: will try and update the spec for merging APIs, that's good news :)
15:40:51 <joadavis> the slides 19 and 20 are showing some gnocchi/aodh and watcher triggering https://www.slideshare.net/GuanYuLin1/autoscale-a-selfhealing-cluster-in-openstack-with-heat
15:41:28 <joadavis> yes, I think the wiki page needs some more information, but the overall method there seemed to work for me (haven't actually triggered autoscaling yet, but the setup worked)
15:41:45 <witek> joadavis: thanks for the references, I think it's a good idea to document using Monasca in that context
15:41:56 <joadavis> agreed
15:42:09 <witek> joadavis: thanks a lot for references
15:42:45 <witek> do we have anything else for today?
15:42:52 <joadavis> (hmm, that might have been mistral not watcher in the slides)
15:43:01 <joadavis> nothing from me
15:44:12 <witek> I will be out of office next week
15:44:18 <witek> and cannot join the meeting
15:44:52 <witek> so please start the meeting without me
15:44:57 <joadavis> will do
15:45:04 <witek> thanks
15:45:27 <witek> I'll be wrapping up
15:45:49 <witek> thank you everyone for joining
15:46:01 <witek> see you in two weeks
15:46:06 <witek> bye, bye
15:46:10 <joadavis> bye
15:46:30 <dougsz> bye all
15:46:36 <witek> #endmeeting