15:00:46 <witek> #startmeeting monasca
15:00:47 <openstack> Meeting started Wed Sep 26 15:00:46 2018 UTC and is due to finish in 60 minutes.  The chair is witek. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:48 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:00:50 <openstack> The meeting name has been set to 'monasca'
15:01:00 <witek> Hello everyone
15:01:24 <koji_n> hello
15:01:39 <witek> hi koji_n, hi pandy
15:01:52 <pandy> Hi witek:)
15:01:55 <joadavis> hi
15:02:01 <witek> hi
15:02:21 <witek> the agenda is pretty full again today
15:02:25 <witek> that's nice :)
15:02:30 <srwilkers> o/
15:02:37 <witek> let's start with leftovers
15:02:42 <witek> https://etherpad.openstack.org/p/monasca-team-meeting-agenda
15:02:53 <witek> #topic pre-upgrade checks
15:03:08 <witek> we have started this actually last week
15:03:43 <witek> there is a new component in oslo for this
15:03:44 <witek> https://github.com/openstack/oslo.upgradecheck
15:03:53 <witek> should we use this one?
15:04:52 <joadavis> I didn't grab the oslo.upgradecheck code because it was still being worked on last week.  It would be nice to use something consistent, but the code is still developing
15:05:17 <joadavis> so I thought it would be good to have a very simple start then refactor it later to use the library when it is available
15:05:44 <dougsz> The library certainly seems very green
15:06:00 <joadavis> Also, the code I wrote will likely need to be moved to a good place for deployment
15:06:29 <witek> Matt has left you some comments in review
15:07:07 <joadavis> yes, I've been working some bugs and haven't incorporated his comments yet, but they were good feedback
15:07:33 <joadavis> s/bugs/unrelated bugs/
15:07:37 <witek> :)
15:07:59 <witek> so, you want to continue to work on this, or prefer to come back to this later?
15:08:23 <joadavis> I don't think it is top priority, but can tinker with it when i have time
15:08:50 <joadavis> If someone else is interested they are welcome to take it up
15:09:15 <witek> ok, thanks
15:09:47 <witek> #topic monasca-notification clean-up
15:10:46 <witek> the first change should probably be the schema update
15:10:57 <dougsz> So we hit some issues configuring plugins via Oslo
15:11:05 <dougsz> Yeah - agreed
15:11:15 <dougsz> It would be nice to push those old patches through
15:11:28 <witek> yes, we need them
15:11:55 <witek> I've seen you've added this to our board
15:11:57 <witek> https://storyboard.openstack.org/#!/board/111
15:12:01 <dougsz> I 've got enough feedback from Johannes to fix up the alembic migration
15:12:09 <witek> great
15:12:17 <dougsz> Thanks to Johannes  + Amir for the review
15:13:06 <dougsz> So that's probably it on this for now - I will focus on the old patches in the little time I have at the moment and circle back around to Charana's ones when those are done.
15:13:42 <witek> is there a particular sequence in which the changes should be merged?
15:13:46 <dougsz> Also, thanks Witek for taking a look!
15:13:59 <dougsz> I will add a 'depends on' to the commit message
15:14:11 <witek> thanks
15:15:11 <witek> btw. during ptg we've agreed to create a second board for tracking bug fixing
15:15:20 <dougsz> ah, ok
15:15:27 <witek> I think the last two changes from your set would be good candidates
15:15:36 <witek> I haven't created the board yet though
15:15:47 <dougsz> Sounds good, I can look at moving it when we have the board
15:15:49 <witek> will do tomorrow
15:16:01 <dougsz> thanks
15:16:27 <witek> thanks Doug for working on this
15:16:34 <witek> #topic new Grafana image
15:16:40 <dougsz> np
15:17:10 <dougsz> pandy?
15:17:20 <pandy> Hi Witek, I have requested two features to be in grafana like getting "OK" & "UNDETERMINED" state in notification and fix "save" in one click
15:17:38 <pandy> it was fixed by Charana & Dougz verified and pushed to master https://github.com/monasca/monasca-grafana/commit/85b7eb46c1cbb561e6c7737b3ee75271fa83af7b
15:17:49 <witek> we have pushed the new image today
15:17:52 <witek> https://hub.docker.com/r/monasca/grafana/
15:18:04 <dougsz> Nice - I thought I saw that fly past my inbox
15:18:11 <pandy> It will be great if we udpate this grafana enhancement to new image, currently, i did rebuild in my own dockerhub repo
15:18:32 <witek> pandy: happened today
15:18:41 <pandy> oh great :)
15:18:55 <witek> https://hub.docker.com/r/monasca/grafana/tags/
15:19:02 <witek> 4.0.0-1.5.1
15:19:45 <pandy> Sure i will check it :) Thats it related to monasca- grafana, let me know can i start my next topic in agend which is lsited at bottom now or will wait ? :)
15:20:13 <witek> let's go in the sequence please
15:20:28 <witek> #topic Stein prios
15:20:47 <witek> I have put together the results of our priorisation game
15:21:28 <witek> as we discussed earlier, I didn't want to put too many tasks
15:21:36 <dougsz> I started writing the Merge APIs spec earlier will have it finished today, then I can add a link to the doc.
15:22:04 <witek> great, I just wanted to ask you folks to add the descriptions :)
15:22:18 <dougsz> :)
15:23:03 <dougsz> Will do.
15:23:09 <witek> also, if you think there is anything important we should add there during the cycle
15:23:17 <witek> we're free to do so
15:23:51 <witek> #topic Kafka config opts in monasca-api
15:24:28 <dougsz> It looks like some refactoring left these behind
15:24:41 <dougsz> Just wanted to check I hadn't missed something before removing them?
15:24:56 <witek> also, that will probably change anyway during the work with confluent-kafka client
15:25:21 <dougsz> Yeah - envisaged some overlap
15:25:53 <witek> I want to finish my Kafka spec this week
15:26:05 <witek> https://review.openstack.org/601268
15:26:20 <witek> so I will go through these options as well
15:26:34 <witek> and will leave you a comment in story
15:26:42 <dougsz> Awesome, thanks witek.
15:27:08 <dougsz> I guess we could backport any small patch to remove them to avoid confusing people with the generated config file from tox
15:28:18 <witek> so you mean, remove all of them now to be in sync with auto-generated, and then backport if needed?
15:28:39 <dougsz> Yeah
15:29:10 <witek> yes, it will work as well
15:30:09 <witek> #topic libvirt upgrade
15:30:52 <witek> joadavis: have you added this one?
15:31:31 <joadavis> yes.
15:31:50 <joadavis> I heard there was a plan for nova to update the libvirt and qemu version for the T release
15:32:05 <joadavis> I just wanted to raise it as a sanity check here - I don't think it will be a problem
15:32:36 <witek> do you have any reference to additional information about this upgrade?
15:33:50 <witek> the problem with libvirt plugin is, that we don't have any automated testing included
15:33:56 <joadavis> there was an email to the openstack-dev mailing list...
15:34:31 <joadavis> Proposal --------  Looking at the DistroSupportMatrix[2], it seems like we can pick the libvirt and QEMU versions supported by the next LTS release of Ubuntu -- 18.04; "Bionic", which are:      libvirt: 4.0.0     QEMU:    2.11
15:35:36 <witek> should we add a task to the backlog for this?
15:36:08 <joadavis> I think the Operating Systems will support it.  It may be good for us to have a task to at least check the release notes for libvirt for any api changes or things that might need an update
15:36:15 <joadavis> in monasca-agent
15:37:19 <witek> also, only partially related, do you know if we could run tempest tests on bionic already?
15:37:53 <joadavis> I don't know.  Are other projects doing that already?
15:38:07 <witek> haven't checked, just came to my mind
15:38:13 <joadavis> good thought
15:39:33 <witek> could you create a testing task for monasca-agent and add it to the board please?
15:39:40 <joadavis> sure
15:39:45 <witek> thanks joadavis
15:40:24 <witek> I'll have a look at running devstack plugin on bionic
15:40:43 <witek> #topic reviews
15:41:27 <witek> rpm-packaging change looks good to me
15:41:48 <joadavis> :thumbsup"
15:42:12 <witek> #topic bugs
15:42:24 <witek> https://storyboard.openstack.org/#!/story/2001593
15:43:06 <witek> to be honest, I haven't looked in detail into it, but don't like the sudoers idea
15:43:50 <witek> it would make the installation/configuration even more complex
15:44:59 <witek> other thoughs?
15:45:08 <joadavis> Yeah, I was struggling with a good answer.  Another engineer here is looking at the problem.  It appears that in the rpm-packaging there are already a few programs set up for monasca-agent user which are granted in the sudoers file
15:46:06 <joadavis> I'm not sure how it works, but if it is possible to grant elevated permission to the monasca-agent user for just the ip command, that might be good.  But as you said, it is a more complex configuration
15:46:44 <joadavis> I wanted to bring it up in case our monasca community has some other great ideas. :)
15:47:33 <witek> we'll ping our team about it
15:47:47 <witek> also, we'll add it to our new bugs board
15:47:54 <joadavis> thanks
15:48:13 <witek> I think we should try to find a nicer solution
15:48:39 <witek> but if that doesn't work, we could set up sudoers
15:49:43 <witek> #topic monasca-agent in large deployments
15:50:18 <witek> pandy: that's yours, right?
15:50:20 <pandy> Hi
15:50:34 <pandy> yes mine
15:50:52 <pandy> quick explanation of my environment
15:50:53 <pandy> Installed monasca using docker on 3 dedicated baremetals, Please note: used monasca docker images for monasca-api, grafana, notification, mysql, influxdb, where kafka & zookeeper used different version
15:50:54 <pandy> for scaling and running on docker swarm mode
15:51:05 <pandy> In 500+ compute installed monasca agent and did configuration using monasca-setup as per https://github.com/openstack/monasca-agent/blob/master/docs/Agent.md
15:51:12 <pandy> Where added libvirt.yaml plugin to collect VM metrics and here configuration of libvirt.yaml  http://paste.openstack.org/show/730940/
15:51:22 <pandy> Issue is some of monasca agent is hitting Neutron API very frequently
15:51:29 <pandy> just suspected  these parameters are causing issue so disabled it
15:51:29 <pandy> ""vm_network_check_enable: false" "ping_check: false" "vm_ping_check_enable: false""
15:51:46 <pandy> Here my question: "Would like to know "Which Monasca agent plugins & parameters will hit Neutron API and whats the interval period ?" & "How can control hitting to Neutron API ?""
15:52:18 <pandy> sorry i was prepared already, so sent messages in row.
15:52:35 <witek> is libvirt the only plugin, you've enabled on compute nodes?
15:53:11 <pandy> by default it has given me process, cpu, network and i have added libvirt.yaml to it
15:54:14 <pandy> default i have congestion, disk, load, network, process, cpu, memory, ntp. I have added "libvirt" into /etc/monasca/agent/conf.d
15:54:41 <pandy> would like to know which plugin is hitting Neutron API ?
15:55:31 <pandy> Neutron.yaml here http://paste.openstack.org/show/730940/
15:55:48 <witek> libvirt plugin communicates with Neutron, but I cannot answer your question about configuration option right now
15:56:14 <pandy> libvirt and ovs plugin only hits neutron
15:56:18 <pandy> as per my understanding
15:56:25 <koji_n> at this moment, ovs and libvirt plugin is using neutronclient
15:56:45 <pandy> i am using libivirt, where disabling "ping_check" will stop hitting neutron API ?
15:57:44 <pandy> FYI, we are not using namespace in compute nodes, so i feel "ping_check" no need, other than that what are parameters inside libivirt will hit Neutron API
15:58:19 <witek> koji_n: will you be able to check this?
15:58:25 <koji_n> https://github.com/openstack/monasca-agent/blob/819822529422e68614dd4c3b95a9b530795e4bcf/monasca_agent/collector/checks_d/libvirt.py#L186-L192
15:58:32 <koji_n> probably, pandy is correct
15:58:52 <koji_n> neutronclient is used only the above section in libvirt plugin
15:59:15 <witek> that was quick, thanks koji_n :)
15:59:39 <pandy> koji_n,  whether "ping_check" only use neutron API  ?
15:59:54 <koji_n> np, but i didn't test it, so maybe it's wrong
16:00:04 <pandy> How about "vm_network_check_enable" ""vm_ping_check_enable: false""
16:00:27 <witek> have to close the meeting
16:00:34 <witek> thanks for joining everyone
16:00:39 <dougsz> Thanks witek, bye all
16:00:39 <joadavis> thanks all
16:00:42 <witek> see you next week
16:00:43 <pandy> witek, i need someone from libvirt to help out this
16:00:55 <witek> #endmeeting