15:00:45 <eglynn> #startmeeting ceilometer
15:00:46 <samuelbercovici> bye
15:00:47 <openstack> Minutes:        http://eavesdrop.openstack.org/meetings/neutron_lbaas/2014/neutron_lbaas.2014-05-08-14.03.html
15:00:47 <nealph> /msg NickServ identify FN49Ford
15:00:49 <openstack> Minutes (text): http://eavesdrop.openstack.org/meetings/neutron_lbaas/2014/neutron_lbaas.2014-05-08-14.03.txt
15:00:49 <german_> bye
15:00:50 <openstack> Log:            http://eavesdrop.openstack.org/meetings/neutron_lbaas/2014/neutron_lbaas.2014-05-08-14.03.log.html
15:00:51 <openstack> Meeting started Thu May  8 15:00:45 2014 UTC and is due to finish in 60 minutes.  The chair is eglynn. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:52 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:00:55 <openstack> The meeting name has been set to 'ceilometer'
15:01:09 <edhall> bye
15:01:17 <eglynn> who's all around for the ceilo meeting?
15:01:20 <ildikov> o/
15:01:22 <llu-laptop> o/
15:01:23 <gordc> o/
15:01:26 <dhellmann> o/
15:01:26 <DinaBelova> o/
15:01:30 <_nadya_> o/
15:01:40 <nsaje> o/
15:01:41 <nealph> o/
15:02:14 <eglynn> cool, so not everyone already on the midnight train to Georgia ;)
15:02:37 <ildikov> eglynn: LOL :)
15:02:38 <enikanorov> \o/
15:02:44 <eglynn> #topic summit finalization
15:02:59 <eglynn> so we had a bit of horse-trading earlier this week around the design track scheduling
15:03:06 <eglynn> (to avoid conflicts with other tracks etc.)
15:03:19 <eglynn> so if you haven't checked the schedule recently, please give it another look ...
15:03:27 <eglynn> #link http://junodesignsummit.sched.org/overview/type/ceilometer
15:03:50 <llu-laptop> is https://wiki.openstack.org/wiki/Summit/Juno/Etherpads#Ceilometer up-to-date?
15:04:07 <eglynn> llu-laptop: yeah good point
15:04:22 <eglynn> anyone who's leading a session, pls prepare an etherpad in advance
15:04:38 <eglynn> and link it to the master list llu-laptop mentions above
15:04:40 <gordc> llu-laptop: i've updated the wiki to reflect latest scheduleing change
15:04:54 <ildikov> sure, will do, thanks for reminding
15:04:55 <eglynn> gordc: thanks much!
15:05:00 <nsaje> Will do
15:05:03 <gordc> eglynn: np
15:05:14 <llu-laptop> gordc: thx
15:05:29 <eglynn> BTW the ceilometer pod will apparently be available from the Monday
15:05:44 <DinaBelova> heh, cool)
15:05:48 <jd__> o/
15:06:10 <eglynn> ... not sure of the exact location of the pod, but the signage should be adequate ;)
15:06:42 <eglynn> we'll be gathering there at 2pm on the Monday to talk about TSDaaS etc.
15:06:45 <ildikov> eglynn: you can never be sure ;)
15:07:13 <eglynn> ildikov: ... yeah I'm bringing my Garmin eTrex just in case ;)
15:07:38 <eglynn> we should also aim to use the pod space for any ad-hoc discussions that come up during the week
15:07:52 <eglynn> ... kinda like a focal point or whatever
15:08:03 <ildikov> eglynn: cool, I will follow you then ;)
15:08:48 <nealph> eglynn: I like that idea...a natural spot to continue discussions "offline"
15:09:18 <eglynn> cool ... I think the pod idea is an experiment for this summit, seems like a good idea to me
15:09:45 <ildikov> eglynn: +1
15:09:51 <eglynn> not sure if there's anything else to discuss about summit?
15:09:55 <DinaBelova> eglynn, possibly it'll be nice to prepare some list of topics? for the pod discussions...
15:09:57 <dhellmann> yeah, the pods take the place of the unconference
15:10:26 <eglynn> DinaBelova: ... yeah I think the scheduling will be fast and loose, first come first served
15:10:28 <DinaBelova> as there were lots of small topics this and last week
15:11:02 <eglynn> DinaBelova: ... maybe just use the fipchart/WB to manage a rough emergent schedule
15:11:16 <DinaBelova> +1
15:11:54 <dhellmann> eglynn: do you know if we're sharing our pod with another team? I know some are, but I don't know how they are grouped.
15:12:26 <eglynn> dhellmann: good question, I heard nova+glance are sharing so likely we'll be too
15:12:59 <eglynn> dhellmann: I'll check ... if I had to guess, my money would be on us sharing with heat
15:13:14 <dhellmann> that would make sense
15:14:17 <eglynn> cool enough, I guess we can move on
15:14:24 <jd__> I hope we share with tempest
15:14:55 <ildikov> jd__: and then with infra too
15:15:00 <eglynn> jd__: yeah that would be good, osmosis by proximity ;)
15:15:17 * jd__ auto-high-five himself for this joke
15:15:29 <eglynn> LOL :)
15:15:34 <ildikov> LOL :)
15:15:45 <eglynn> #topic BP review process
15:15:57 <eglynn> I took an action in our previous meeting to bring a proposal to the table on that
15:16:00 <eglynn> so here goes ...
15:16:21 <eglynn> ... the idea is mainly to follow the nova lead on this, but with some tweaks
15:16:46 <eglynn> first I don't think there's a need to have a separate team responsible for BP-review and code-review
15:17:05 <eglynn> so the ceilo-drivers team should remain equivalent to the ceilo-core team IMO
15:17:12 <eglynn> #link https://launchpad.net/~ceilometer-drivers/+members#active
15:17:24 <eglynn> (whereas for nova AFAICS, only a subset of the cores ...
15:17:34 <eglynn> https://review.openstack.org/#/admin/groups/25,members
15:17:40 <eglynn> ... are also drivers:
15:17:47 <eglynn> https://launchpad.net/~nova-drivers/+members#active )
15:18:04 <eglynn> that distinction doesn't seem necessary to me
15:18:22 <ildikov> I think currently we can handle it and later we can decide to have a separate team for this, if needed
15:18:35 <eglynn> yeap ... on a project of the size of ceilo, one team of reviewers is plenty
15:19:05 <eglynn> k, that's uncontroversial so
15:19:22 <eglynn> second, I hacked out a draft BP template
15:19:27 <eglynn> based on nova but customized for ceilo-specifics
15:19:37 <eglynn> #link https://etherpad.openstack.org/p/ceilometer-juno-blueprint-template
15:19:57 <eglynn> ... actually, easier to read as exported txt
15:20:02 <eglynn> https://etherpad.openstack.org/p/ceilometer-juno-blueprint-template/export/txt
15:20:45 <eglynn> as mentioned on the chanell earlier ...
15:20:51 <eglynn> TL;DR: the explicit diffs WRT the nova original are ... http://fpaste.org/100125/54945513/
15:22:24 <eglynn> ... dunno if there's anything contraversial in that
15:22:37 <eglynn> ... it's mainly motherhood and applepie ;)
15:23:03 <eglynn> other than maybe: The ceilometer project is explicitly not interested in "code drops", ...
15:23:10 <ildikov> eglynn: I'm not sure we should drop unit/scenario tests and focus only on tempest
15:23:25 <ildikov> eglynn: but this topic can be further discussed on the testing session in ATL
15:23:46 <eglynn> ildikov: cool ... feel free to hack away on that etherpad with your thoughts
15:24:08 <_nadya_> ildikov: absolutely. tempest is not for unit testing
15:24:29 <ildikov> eglynn: cool, I will do that before the session and include this etherpad in the session agenda too
15:25:05 <eglynn> _nadya_: so the point in the BP template is that unit/scenario test coverage is assumed, so don't bother telling us about it
15:25:24 <ildikov> eglynn: coverage is not enough
15:25:35 <eglynn> _nadya_: ... whereas we *really* want to know about your plan tempest coverage! :)
15:25:44 <ildikov> eglynn: we need more negative tests for instance, than we have now
15:26:24 <_nadya_> eglynn: we will discuss it on topic about tempest :)
15:27:05 <ildikov> eglynn: but we can move on now and discuss it on the summit, we have several topics for today
15:27:25 <eglynn> anyhoo the idea would be that any new BPs for Juno would be proposed to gerrit, along the lines of the nova-specs project ...
15:27:39 <eglynn> https://review.openstack.org/#/q/status:open+project:openstack/nova-specs,n,z
15:28:01 <eglynn> anyone know what's required to set up a new repo under https://github.com/openstack ?
15:28:07 <ildikov> eglynn: so this means that the currently proposed ones should be rewritten according to the template, right?
15:28:35 <llu-laptop> eglynn: just want to ask the same question about ceilometer-specs
15:28:39 <eglynn> ildikov: ... yes anything targetted at Juno should go thru the same process
15:28:47 <eglynn> (IMO)
15:28:58 <ildikov> eglynn: do we have a deadline for this?
15:28:59 <gordc> ildikov: good way to filter out which of the current bps are actually active.
15:29:05 <dhellmann> eglynn: I can help with the repo setup
15:29:22 <eglynn> dhellmann: excellent! ... I was hoping someone would step up :)
15:29:23 <dhellmann> eglynn: https://review.openstack.org/92614
15:29:30 <DinaBelova> I guess it
15:29:40 <DinaBelova> will be needed to make change to the infra-config
15:29:42 <dhellmann> eglynn: if you give me a github repo that you want imported, I can shepherd the change through
15:29:49 <nealph> gordc: +1
15:29:49 <ildikov> gordc: sure, I agree, I just need to know that until when, I should rewrite my owns :)
15:30:00 <eglynn> dhellmann: cool, will do, thank you sir!
15:30:21 <DinaBelova> eglynn, dhellmann- smth like this? https://review.openstack.org/#/c/91723/
15:30:42 <eglynn> ... so all this is just a proposal that I'm seeking buy-in on from the project team
15:30:57 <eglynn> ... does anyone have any reservations about the idea?
15:31:21 <gordc> none from me.
15:31:26 <llu-laptop> I think we also need to update the CM section around https://wiki.openstack.org/wiki/Blueprints#Nova
15:31:40 <eglynn> gordc: cool :)
15:31:45 <eglynn> llu-laptop: agreed
15:32:04 <ildikov> I like the idea now
15:32:23 <ildikov> I will less like it during the refactor process and after that I will like it again ;)
15:32:49 <eglynn> ildikov: LOL :) ... think of it an "investment"
15:32:59 <dhellmann> DinaBelova: yes
15:33:08 <eglynn> (hard to think of word-smithing as productive work, I know ...)
15:33:30 <eglynn> so once we've the repo and gerrit project set up, I guess we could review the BP template on gerrit
15:33:47 <eglynn> (using the content in the etherpad linked above as the initial version)
15:34:02 <DinaBelova> eglynn, cool)
15:34:49 <eglynn> k, sounds like no need to put all that new process overhead to a vote
15:34:58 <ildikov> eglynn: cool, I guess we will have more BP after the summit, than before, so it is ok to use that as an initial
15:35:08 <eglynn> cool
15:35:09 <ildikov> eglynn: and also prolly it will not change much
15:35:27 <eglynn> yeap agreed
15:35:41 <eglynn> k, best move on I guess?
15:35:48 <eglynn> #topic tempest status
15:36:28 <_nadya_> we've started working on scenario tests
15:36:38 <eglynn> cool
15:37:09 <_nadya_> #link https://review.openstack.org/#/c/92108/
15:37:26 <eglynn> _nadya_: BTW thanks for the tempest input in ildikov's test strategy etherpad for summit
15:38:01 <_nadya_> yep, it's just a draft and any other inputs are very welcome
15:38:19 <_nadya_> JFYI #link https://etherpad.openstack.org/p/ceilometer-test-strategy
15:38:51 <eglynn> are we overloading the term "scenario test" a bit here?
15:38:54 <eglynn> ... but I guess that's the usual Tempest terminology
15:39:17 <_nadya_> and that's all from me. btw, does anyone know the status of ubuntu14 on gating?
15:39:33 <ildikov> _nadya_: thanks, my notebook is half dead, so I got gibi's for the time of the meeting... :S
15:39:49 <_nadya_> eglynn: afaik "scenario" is tempest term for more compex tests
15:39:50 <eglynn> _nadya_: haven't heard anything, but I intended to twist some infra arms in ATL next week
15:40:39 <eglynn> _nadya_: cool, makes sense ... just contrasting with our mongo/sqla/hbase/db2 scenario tests in the ceil code-tree
15:40:46 <_nadya_> eglynn: yep, it's great opportunity
15:40:47 <eglynn> *ceilo
15:41:01 <eglynn> cool, I guess we can move on?
15:41:05 <_nadya_> yep
15:41:17 <eglynn> #topic Monitoring-as-a-Service
15:41:31 <eglynn> raised by aviau on the ML
15:41:37 <aviau> Hello everyone!
15:41:40 <eglynn> #link http://lists.openstack.org/pipermail/openstack-dev/2014-May/034189.html
15:41:49 <aviau> I have recently submitted a blueprint to Ceilometer and I wanted to get in touch with you. We are looking for reactions from the Ceilometer team for a possible Monitoring service.
15:41:51 <eglynn> aviau: hey! :)
15:42:00 <eglynn> aviau pity you're not going to be in ATL next week to discuss face-to-face
15:42:04 <aviau> #link https://blueprints.launchpad.net/ceilometer/+spec/monitoring-as-a-service
15:42:07 <ildikov> aviau: hi
15:42:13 <eglynn> aviau: ... but the floor is your's for now :)
15:42:27 <aviau> eglynn, I will be available for the next summit.
15:42:44 <eglynn> aviau: cool ... how do you want to move the discussoin forward in the meantime?
15:42:53 <eglynn> aviau: ... a recurring topic in this meeting maybe?
15:43:18 <aviau> eglynn, This would be a great idea.
15:43:34 <eglynn> aviau: cool, let's aim to that starting the week after summit
15:43:43 <eglynn> (no meeting next week)
15:44:05 <aviau> eglynn, Good, I will hold discussions until that meeting.
15:44:11 <eglynn> aviau: cool
15:44:41 <eglynn> BTW we do need to be cognizant of the perception of "misson creep"
15:44:56 <eglynn> ... as expressed say in the debate around the project mission statement
15:45:06 <eglynn> #link https://review.openstack.org/87526
15:45:35 <eglynn> but overall I think we shouldn't be too afraid to explore this direction
15:45:42 <aviau> What is your take on this?
15:46:14 <eglynn> aviau: my gut feeling is "big tent" :)
15:46:37 <aviau> Hopefully that will work. I think that Monitoring is closely related to Metering.
15:46:45 <ildikov> eglynn: do you mean to explore the monitoring direction?
15:47:08 <gordc> aviau: i would think your bp falls under the mission statment we had since grizzly (whether any work was done on it or not)
15:47:27 <ildikov> aviau: do you plan to implement a monitoring tool-like service or it is focusing more on events in OpenStack?
15:47:35 <eglynn> ildikov: yes, I think we should at least be open to that possibility
15:47:56 <aviau> ildikov, For how, the focus is not on the events in OpenStack.
15:48:31 <aviau> However, this is still a blueprint. Now work has been done yet and we hope to receive feedback.
15:48:36 <aviau> no*
15:49:03 <ildikov> aviau: I meant that monitoring means to ping the hosts/guests to check availability or processing more kinds of events coming from the other services than we deal with currently?
15:49:10 <gordc> aviau: seems like it's a meld of polling agents and events -- i'll make comments to your bp.
15:49:46 <eglynn> seem like this BP would be a good guinea pig for the BP review process?
15:50:21 <ildikov> eglynn: this seems to be an excellent candidate :)
15:50:23 <aviau> ildikov, ping the host/guests to check availability. It would be great if we supported Nagios standards.
15:51:08 <aviau> I will read up on the review process, thanks.
15:51:38 <eglynn> aviau: ... ^^^ discussed above (mechanics not yet in place)
15:51:42 <_nadya_> aviau: I'm not nagios expert but can use Nagios+Ceilo together?
15:51:49 <ildikov> aviau: do we plan to do this, when there are n+1 monitoring tools that can do that?
15:52:17 <_nadya_> aviau: I mean not to create a Nagios-like-tool but use Nagios itself?
15:52:28 <aviau> #link https://github.com/savoirfairelinux/check_ceilometer
15:52:33 <aviau> #link http://blog.zhaw.ch/icclab/nagios-ceilometer-integration-new-plugin-available/
15:52:41 <aviau> You can use nagios to monitor OpenStack.
15:53:27 <_nadya_> aviau: so why it may be useful to create one more Nagios inside Ceilo?
15:53:44 <_nadya_> aviau: I'm just asking, to understand better :)
15:53:50 <ildikov> aviau: and what would MaaS do then?
15:54:08 <eglynn> ... sounds like there's some scoping needed here, to avoid the perception of wheel-reinvention
15:54:08 <aviau> however, Ceilometer currently won't give you all the information you need about your guests. Ceilometer does not know if my web service is up, for example.
15:54:36 <aviau> eglynn, Absolutely! Work has to be done on possible use cases.
15:55:11 <ildikov> aviau: I'm not sure Ceilometer should check web services
15:55:22 <eglynn> aviau: ... so we need a justification for why ceilo should be concerned about that, as opposed to just using existing tooling
15:55:54 <eglynn> aviau: (not necessarily needed right now, but food for thought and future discussoin)
15:56:23 <eglynn> ... k, let's continue the discussion on gerrit?
15:56:29 <eglynn> (... once the git/gerrit incantations have been incanted)
15:56:30 <_nadya_> yep
15:56:37 <aviau> Yep
15:56:47 <eglynn> #topic Issues raised by https://review.openstack.org/#/c/86408/, how to preserve per disk metrics, per net interface metrics?
15:56:49 <ildikov> yep, we are running out of time here
15:57:14 <eglynn> prad__: defer that discussion ^^^ to ATL?
15:57:20 <ildikov> will we have a chance to discuss this on the summit?
15:57:33 <prad__> eglynn: sure i’m cool with that
15:57:39 <_nadya_> mm :)
15:57:52 <eglynn> prad__: one for the pod methinks
15:58:00 <prad__> probably a good topic for pod
15:58:16 <DinaBelova> prad__ +1
15:58:16 <_nadya_> I will try to summarize all my thoughts too
15:58:22 <ildikov> eglynn, prad__: +1
15:58:32 <llu-laptop> +1 for pod
15:58:42 <_nadya_> *somewhere, in cr maybe
15:58:53 <prad__> thx _nadya_
15:59:06 <prad__> does it make sense to have an etherpad for pod with topics to discuss
15:59:08 <eglynn> TBH the sample-datapoints-in-resource-metadata idea will be problematic in the context of TSDaaS
15:59:13 <eglynn> ... but let's punt to the pod
15:59:21 <ildikov> prad__: I just wanted to ask you to prepare one :)
15:59:32 <prad__> sure i can do that
15:59:40 <eglynn> 30 second warning ;)
15:59:42 <ildikov> prad__: cool, thanks
15:59:43 <_nadya_> as for me it's related to data-model topic too
15:59:52 <eglynn> _nadya_: yeap
15:59:57 <eglynn> #topic open discussion
16:00:16 <eglynn> up against the shot-clock
16:00:35 <eglynn> anything else maybe we bring to the project channel?
16:00:45 <DinaBelova> +1
16:01:05 <eglynn> thanks as always folks for a productive meeting!
16:01:07 <eglynn> #endmeeting ceilometer