15:01:54 <eglynn> #startmeeting ceilometer
15:01:55 <openstack> Meeting started Thu Jun 26 15:01:54 2014 UTC and is due to finish in 60 minutes.  The chair is eglynn. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:01:56 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:01:58 <openstack> The meeting name has been set to 'ceilometer'
15:02:00 <jd__> o/
15:02:01 <eglynn> hey folks
15:02:02 <Alexei_987> o/
15:02:02 <DinaBelova> o/
15:02:03 <shengjiemin> o/
15:02:06 <llu-laptop> o/
15:02:07 <vrovachev> o/
15:02:08 <idegtiarov> o/
15:02:09 <gordc> o/
15:02:10 <ityaptin> o/
15:02:10 <ildikov> o/
15:02:11 <nealph> o/
15:02:15 <KurtRao> o/
15:02:26 <enikanorov_> adolf would be pleased to see this
15:02:27 <eglynn> #topic Juno-2 planning, progress on blueprint filing
15:02:51 <fabio_> o/
15:02:58 <eglynn> so you probably saw the ML thread on a cut-off point for spec filing and landing
15:03:18 <llu-laptop> same day as nova?
15:03:19 <eglynn> it involves a bunch of new TLAs
15:03:28 <eglynn> llu-laptop: yeah, that was the idea
15:03:31 <tongli> o/
15:03:42 <llu-laptop> jul 4th & jul 10th?
15:03:43 <eglynn> #link https://review.openstack.org/#/q/project:openstack/ceilometer-specs,n,z
15:04:04 <eglynn> llu-laptop: +/- a day, I was thinking more July 11th (end of week)
15:04:29 <llu-laptop> ok
15:04:44 <eglynn> so I guess the message is ... if you have a spec you want to propose for juno, don't delay!
15:05:00 <llu-laptop> i have to hurry up :(
15:05:00 <eglynn> ... oh, dunno, maybe for SNMP improvements? ;)
15:05:08 <eglynn> LOL :)
15:05:25 <eglynn> how about we copy the cool kids in Ironic?
15:05:30 <shengjiemin> does that mean all the reviews need to be finished by them?
15:05:38 <shengjiemin> i meant all the specs
15:05:43 <eglynn> ... and do a "review jam" together while in Paris?
15:05:54 <DinaBelova> eglynn +1
15:06:00 <fabio_> eglynn +1
15:06:02 <ildikov> eglynn: that could work
15:06:22 <eglynn> shengjiemin: SPD ... deadline for spec proposal Jul 4th, SAD .. deadline for spec landing Jul 11th
15:07:11 <eglynn> piyush had a spec in the works on declarative metering, but no sign yet
15:07:30 <shengjiemin> eglynn: ok
15:07:42 <eglynn> llu-laptop: otherwise it might make sense to re-purpose that idea ^^^ for code-free SNMP metering?
15:08:35 <llu-laptop> I don't think they're alike,
15:09:07 <eglynn> llu-laptop: TBH the declarative metering idea seems even more useful for SNMP (to me)
15:09:29 <eglynn> llu-laptop: ... i.e. a generic SNMP OID -> (meter name, unit, type) mapping
15:10:19 <eglynn> llu-laptop: ... anyway just a thought, lsmola__ would be a very happy man if such a thing were to exist for SNMP
15:10:28 <llu-laptop> snmp has something different than the hyperv's. but snmp have some metrics which can not directly mapped from OIDs.
15:11:03 <eglynn> llu-laptop: a-ha ok, I guess we should discuss the details off-line
15:11:11 <llu-laptop> I'll try to leverage that kind of thought as much as I could, at least for those snmp oids which can be directed map to metrics
15:11:33 <eglynn> llu-laptop: that would be great! :) ... thank you sir!
15:12:18 <lsmola__> llu-laptop: I agree, it will be hard to define it in general
15:12:33 <eglynn> k, not much more to say on juno-2 planning other than the obvious
15:12:37 <eglynn> ... i.e. the clock is a-tickin'!
15:12:43 <lsmola__> llu-laptop: btw. it should be somehow tied to discovery and snmp
15:13:25 <eglynn> move on, or?
15:13:34 <llu-laptop> lsmola__: yes, it will depend on your patch
15:13:37 <llu-laptop> go on
15:13:39 <eglynn> #topic TSDaaS/gnocchi update
15:13:49 * jd__ clears his throat
15:13:54 <eglynn> jd__: ... the floor is your's :)
15:13:55 <lsmola__> llu-laptop: so for some hardware we can have group of OIDs SNMPv3, for e.g. some switches it would be other group of OIDs etc..
15:13:56 <jd__> or fingers
15:14:17 <jd__> not a lot of update on the code actually this week
15:14:28 <jd__> but we exchanged too many emails on how to do a proper aggregation
15:14:41 <jd__> I think we had a lot of progress on this area
15:15:10 <jd__> now I should be able to write some code, very soon
15:15:21 <lsmola__> llu-laptop: btw. I noticed that if you have to do SNMP walk through multiple devices, you store the device info under same resoruce but different metadata
15:15:22 <eglynn> yeah I think we're slowly converging, apols for labouring the point to death
15:15:27 <jd__> maybe not today because my brain's still burning :-)
15:15:49 * DinaBelova thinking eglynn started one more mailing thread to kill jd__ :D
15:16:00 <lsmola__> llu-laptop: is there a way to retrieve list of devices in resource? and e.g. stats for only one device?
15:16:12 <DinaBelova> jd__, btw, what's about the https://review.openstack.org/#/c/98798/ ? are you planning to continue working on it, or it should be chosen someone to continue it?
15:16:32 <lsmola__> llu-laptop: I didn't check for a while what operations with metadata are allowed
15:17:02 <jd__> DinaBelova: I will, but I prefer to work on the aggregation stuff we're discussing with eglynn first
15:17:08 <jd__> so if someone wants to step in, be my guest
15:17:13 <DinaBelova> jd__, ok :)
15:17:26 <DinaBelova> I guess it's a good idea while the aggregation ideas are fresh and hot
15:17:27 <eglynn> yeah I think we should get the aggregation and tagging tied down first
15:17:50 <jd__> yeah that's why I keep reading and sending emails :)
15:17:59 <eglynn> on those long-winded email threads, jd__ would you prefer on dev-list discussion or OK with a smaller group?
15:18:00 <DinaBelova> jd__, :D
15:18:15 * eglynn tried to create a openstack-gnocchi list, didn't fly tho'
15:18:24 <jd__> eglynn: I'm ok with a smaller group, it's already hard enough to follow
15:18:33 <eglynn> jd__: cool enough
15:18:46 <jd__> having random smart ass poping in with smart idea or URLs isn't gonna help me
15:18:56 <DinaBelova> eglynn, as for the tagging - I've went through jd__'s answer about the autoscaling, really I feel tagging is still useful, but I see no clear objections for the his solution too
15:19:04 <jd__> (ok sometimes it does but it's rare :-))
15:19:21 <eglynn> yeah low noise-to-signal ration sometimes
15:19:26 <eglynn> *ratio
15:19:38 <jd__> (eglynn: though I'd love to have a openstack-ceilometer list ultimately, but life…)
15:19:54 <eglynn> DinaBelova: yeah, I need to digest jd__'s response on that tagging question
15:19:56 <jd__> "ah, people" ya know.
15:20:09 <jd__> cool
15:20:20 <eglynn> jd__: yeah, me too, but I think the os-infra are against per-project lists in general
15:20:51 <eglynn> I guess we could set up something external, say googlegroups or the likes ...
15:21:03 <DinaBelova> probably we may go to the open discussion?
15:21:08 <jd__> I hear revolution whispered
15:21:15 <eglynn> LOL :)
15:21:16 <DinaBelova> :D:D:D
15:21:30 <DinaBelova> that's open discussion anyway :D
15:21:40 <eglynn> ... before open discussion, let's talk about gay Paree!
15:21:44 <eglynn> #topic Mid-cycle agenda
15:22:01 <DinaBelova> eglynn, hehe, when I looked on the agenda it was no this item there :D
15:22:03 <DinaBelova> sorry
15:22:18 <jd__> gay Paree, lol
15:22:19 <eglynn> here's an idea, how about an etherpad for a super-lightweight agenda
15:22:27 <DinaBelova> eglynn, +1
15:22:32 <eglynn> DinaBelova: added it a few minutes ago
15:22:42 <jd__> good idea
15:22:50 <eglynn> so maybe just a rough list of topics with owners
15:22:54 <ildikov> eglynn: like go there, talk and then come home? ;)
15:23:09 <SergeyLukjanov> eglynn, we've been using etherpad for managing storyboard sprint agenda this winter and it was working well
15:23:28 <DinaBelova> eglynn, https://etherpad.openstack.org/p/paris-July2014-ceilometer-sprint ?
15:23:39 <eglynn> ildikov: nope more come prepared with research done on the topic and ready to finalize the discussion and write some code!
15:24:26 <ildikov> eglynn: very ambitious ;)
15:24:30 <eglynn> DinaBelova: cool! ... I had some v. rough notes, pasting
15:25:14 <eglynn> ... if folks could pile in on DinaBelova's etherpad and start adding/subtracting topics, that would be great!
15:25:55 <gordc> i'm still working on approval but if i do make it i should have some info influx... playing around with it right now.
15:26:05 <eglynn> ... and the agenda writes itself, /me loves it! :)
15:26:29 <eglynn> gordc: cool, fingers crossed the approval will come through!
15:26:55 * jd__ . o O (if only the code could do that too and writes itself)
15:27:07 <DinaBelova> eglynn, jd__ :D
15:27:31 <DinaBelova> #link https://etherpad.openstack.org/p/paris-July2014-ceilometer-sprint
15:27:31 <eglynn> jd__: LOL :) ... what, you tellin' me there's no "generate code" button on powerpoint!?!! ;)
15:27:38 <jd__> hehe
15:28:07 <eglynn> jd__: anything else on logistics, e.g. start time on Wednesday morning?
15:28:51 <ildikov> eglynn: jd__: next cycle's innovation can be model driven development in Ceilo, so you can draw the code instead of writing :)
15:29:12 <DinaBelova> ildikov, I guess we'll have only girl-team if so
15:29:27 <DinaBelova> ildikov - who love drawing :)
15:29:53 <eglynn> ildikov: email-driven codegen is what /me needs ;)
15:30:01 <ildikov> DinaBelova: I love, but with a pen on real paper and not really UML diagrams, but well, maybe you're right :)
15:30:22 <eglynn> I guess we're in open discussion territory already ...
15:30:27 <ildikov> eglynn: I don't have that much capacity for reviewing ;)
15:30:29 <eglynn> #topic open discussion
15:30:41 <DinaBelova> ok, so I'd love to ask eglynn and amalagon :)
15:30:46 <Alexei_987> I would like to make a small announcement
15:30:47 <DinaBelova> about that mystery docco
15:30:54 <DinaBelova> Alexei_987, oh, ok, go ahead
15:30:54 <Alexei_987> I have good and bad news for you
15:30:56 <amalagon> err...
15:31:09 <Alexei_987> and the news is that I will no longer contribute to ceilometer
15:31:24 <Alexei_987> so I wish you all good luck
15:31:40 <Alexei_987> and my farewell words could be “I don’t know half of you half as well as I should like; and I like less than half of you half as well as you deserve.”
15:31:47 <Alexei_987> thanks for your patience :)
15:31:55 <eglynn> Alexei_987: good luck and thank you for your contributions
15:32:07 <DinaBelova> Alexei_987, thank you for taking part :)
15:32:07 <eglynn> Alexei_987: ... are you staying around the community on other projects?
15:32:26 <Alexei_987> eglynn: yes I will contribute to other openstack projects
15:32:49 <eglynn> Alexei_987: ok, well maybe our paths will cross again in the future in that case ...
15:33:02 <ildikov> Alexei_987: all good to you too
15:33:05 <Alexei_987> eglynn: sure
15:33:24 <gordc> Alexei_987: good luck! pop back in some times. your reviews and contributes were very useful.
15:33:24 <llu-laptop> Alexei_987: good luck, sorry to lose you in CM
15:33:54 <DinaBelova> amalagon, eglynn - returning back to my question...
15:33:58 <shengjiemin> Alexei_987: good luck
15:34:01 <eglynn> DinaBelova: ... so that write-up that Ana is working on
15:34:05 <amalagon> oh yeah: mystery docco = https://docs.google.com/document/d/1VUk24Ft6SDYuQX9cAL12Dmz1AFQYIaO6qebC93h8xHY/edit#
15:34:06 <eglynn> DinaBelova: (options for computing moving-averages, exponential smoothing and the likes, over gnocchi time-series)
15:34:10 <DinaBelova> yes, yes :)
15:34:23 <DinaBelova> amalagon, thanks!
15:34:37 <amalagon> sure!
15:34:39 <DinaBelova> wow-wow, I see eglynn was there :D
15:35:23 <DinaBelova> amalagon, I'll go through it, as that's really interesting topic
15:35:39 <DinaBelova> amalagon, won't you mind if I'll post some comment if I'll have some?
15:35:45 <amalagon> please do!
15:35:57 <DinaBelova> amalagon, np :)
15:36:27 <DinaBelova> amalagon, that was some kind of this I wanted to bring up to the open disc :)
15:36:38 <DinaBelova> not to flood all other parts
15:36:56 <eglynn> DinaBelova: so no mystery was intended, the doc was just a way of gathering together some IRC brainstorming in a more coherent form
15:37:28 <eglynn> ... also good to have a record of progress on OPW work
15:37:30 <DinaBelova> eglynn, yeah, I just used this word as nobody could see that before :)
15:37:40 <DinaBelova> except u & Ana
15:37:42 <DinaBelova> :)
15:37:55 <DinaBelova> eglynn, do we have other things to discuss here?
15:38:04 <eglynn> BTW amalagon discovered an interesting bug in the etherpad-lite version used in openstack
15:38:11 <eglynn> (hence the use of googledocs)
15:38:39 <eglynn> TL;DR: don't rely on openstack etherpad to store long docs
15:38:42 <DinaBelova> eglynn, hm, which one?
15:38:44 <DinaBelova> oh
15:38:46 <amalagon> it froze but eglynn saved the day - /export/txt
15:38:51 <llu-laptop> what's the bug, amalagon
15:38:58 <DinaBelova> oh, that's ugly...
15:39:06 <amalagon> the ether pad just hung on "loading" and wouldn't show the content
15:39:08 <eglynn> there's a magic number of lines, maybe 800, after which it dies
15:39:15 <DinaBelova> thanks for naminmg this issue...
15:39:38 <ildikov> amalagon: it usually gets disconnected me after a few mins and then some of my typing is gone, so I never liked it that much...
15:40:06 <eglynn> ... the garish line colours make my eyes bleed ;)
15:40:30 <amalagon> :)) /me big fan of the garish line colors
15:41:07 <llu-laptop> amalagon: thanks for the findings, we some quite some meeting minutes on etherpad and it's growing, i'll save them somewhere else
15:41:07 <ildikov> eglynn: girls like it :)
15:41:19 <llu-laptop> ildikov: :)
15:41:30 <amalagon> llu-laptop, good call
15:41:35 <eglynn> ildikov: any colour is fine with /me ... as long as it's black ;)
15:41:54 <amalagon> lol
15:43:21 <eglynn> anything else folks?
15:43:24 <shengjiemin> quick one - guys, as I mentioned last week, added in a zabbix bp-specs: https://review.openstack.org/#/c/102713/
15:43:33 <shengjiemin> pls have a look :)
15:43:39 <DinaBelova> eglynn, nope
15:43:49 <eglynn> shengjiemin: will do
15:43:57 <shengjiemin> eglynn: cheers
15:43:57 <ildikov> eglynn: we had a balck Ford right after I had my driving license, that was my favourite car ever with the traditional color :)
15:44:29 <eglynn> very appropriate, given the originator of that saying!
15:44:36 <eglynn> ok
15:44:40 <eglynn> ... for those coming to Paris next week, happy travels!
15:44:49 <eglynn> ... hopefully the air traffic controllers won't go on strike
15:44:52 <eglynn> ... again! ;)
15:44:54 <shengjiemin> eglynn: u guys have fun!!!
15:45:03 <ildikov> and what time finally then?
15:45:04 <shengjiemin> eglynn: LOL
15:45:21 <eglynn> ildikov: dunno, 9am?
15:45:42 <eglynn> ... or is earlier traditional in France?
15:45:55 <eglynn> jd__: ^^^
15:46:07 <ildikov> eglynn: I'm fine with 9am :)
15:46:11 <eglynn> cool
15:46:45 <eglynn> see y'all next week then ... or failing that, next November same place! :)
15:46:51 <eglynn> #endmeeting ceilometer