15:00:27 <gordc> #startmeeting ceilometer
15:00:29 <openstack> Meeting started Thu Aug 13 15:00:27 2015 UTC and is due to finish in 60 minutes.  The chair is gordc. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:30 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:00:32 <openstack> The meeting name has been set to 'ceilometer'
15:00:36 <cdent> word
15:01:25 <ildikov> o/
15:01:41 <llu-laptop> o/
15:02:04 <eglynn> o/
15:02:08 * gordc turns off music
15:02:10 <idegtiarov> o/
15:02:56 <nadya> o/
15:03:02 <gordc> let's go!
15:03:06 <fabiog> hi
15:03:10 <gordc> self.assertRaises(webtest.app.AppError,
15:03:10 <gordc> self.get_json,
15:03:10 <gordc> '/samples?limit=-2')
15:03:13 <gordc> arg
15:03:23 <gordc> #topic recurring: roadmap items (new/old/blockers) https://wiki.openstack.org/wiki/Ceilometer/RoadMap
15:03:44 <gordc> seems like we're making progress on remaining bps
15:03:59 <gordc> functional tests are gated now
15:04:16 <gordc> mandatory limits is almost done
15:04:43 <gordc> prads making progress on declarative meters
15:04:55 <gordc> llu-laptop: one question i had was the snmp spec
15:05:01 <gordc> llu-laptop: is that being worked on still?
15:05:09 <jd__> hoy
15:05:17 <llu-laptop> yes, i'm working on that, but progress is not fast
15:05:19 <llu-laptop> sorry
15:05:32 <gordc> llu-laptop: no worries... just checking
15:05:52 <gordc> we have roughly 2.5 weeks... you think it's still doable?
15:06:11 <llu-laptop> i'll try my best
15:06:35 <gordc> keeping in mind the gate usually explodes the last two weeks
15:06:55 <gordc> llu-laptop: ok. let me know if we need to push it i guess... or if we need to request extension
15:07:35 <llu-laptop> gordc: ok, got that
15:08:05 <gordc> jd__: any idea regarding aodh? did sileht ever get it all gated?
15:08:15 <gordc> llu-laptop: cool cool
15:08:41 * gordc is assuming sileht is super busy or on pto
15:08:42 <cdent> gordc: sileht was adding integration tests, I've been doing the functional tests and they are merged
15:08:48 <jd__> gordc: sileht is on PTO this week
15:08:59 <cdent> so aodh gates on some (quite lightweight) functional tests
15:08:59 <gordc> jd__: ack
15:09:08 <jd__> gordc: what part being gated are you talking about? cdent did some of that
15:09:18 <gordc> cdent: cool stuff
15:09:43 <gordc> jd__: mainly on the integration between ceilo+aodh+heat... that's probably the main use case we need to worry about
15:10:09 <gordc> once we have that, ildikov can start drafting the docs on difference with aodh
15:10:10 <cdent> those are currently not working
15:10:11 <jd__> not sure if we have that complete yet but cdent and sileht are on it :]
15:10:25 <cdent> but they just need some tweaks
15:10:27 <ildikov> gordc: +1
15:10:42 <jasonamyers> o/
15:10:44 <gordc> cdent: cool cool. nice to know there aren't any blockers
15:11:40 <gordc> those are the two main items i was tracking... any one else with specs they want to bring up?
15:12:45 <gordc> cdent: you debating whether or not you want to gnocchi dispatcher cache?
15:12:57 <gordc> to do*
15:13:08 <cdent> I'm waiting to see what magic jd__ brings when he fixes the pending dispatcher bug
15:13:29 <gordc> cdent: cool cool. we can revist in gnocchi topic then.
15:13:30 <cdent> if that makes it clean, then I will likely revisit it
15:13:35 <ildikov> gordc: IIRC we still have a few open patches for event-alarms as for BPs
15:13:44 <jd__> 💫 ✨
15:13:54 * gordc see squares
15:13:57 <gordc> or eye?
15:14:21 <gordc> ildikov: yeah. i think it's better now that we have migration support so we can start making progress in aodh
15:14:30 <jd__> gordc: go get an IRC client with Unicode support :p
15:14:47 * gordc doesn't understand technology
15:15:17 <gordc> ildikov: i think because aodh is running on it's own release schedule, we're less tied to time
15:15:35 <gordc> but definitely, start reviewing Ryota's alarm patches.
15:15:55 <ildikov> gordc: when do we plan to release it?
15:16:14 <gordc> ildikov: good question.
15:16:21 <jd__> at least at Liberty time
15:16:31 <jd__> I'd hope to make a first 1.0.0 by the end of the month
15:16:48 <jd__> but I'd like to clean some stuff still
15:16:51 <gordc> ildikov: i'd probably say 1.0.0 could be when we have equivalence to ceilometer-alarm
15:17:12 <gordc> i don't think we need to wait for event-alarms to release
15:17:27 <gordc> or whatever jd__ said ^^
15:17:53 <ildikov> it would be good to keep that feature in release timeframe if it's ready
15:18:15 <gordc> ildikov: agreed.
15:18:55 <jd__> gordc: what's equivalence?
15:19:56 <gordc> jd__: the integration gate stuff... basically same functionality from user pov
15:20:13 <jd__> AFAIK we have that from the beginning no?
15:20:15 <gordc> same or more*
15:20:19 <jd__> "same functionnalities"
15:20:31 <jd__> we just don't have the proof :)
15:20:36 <gordc> jd__: yeah. so basically gate... :)
15:21:02 <gordc> just want to avoid having a broken 1.0.0
15:21:52 <gordc> we all good?
15:22:01 <gordc> regarding new specs...
15:22:52 <gordc> i'm not declaring a freeze... but tbh, new specs will have lower priority and depending on how complicated it is, probably won't survive the end of cycle gate
15:23:23 <gordc> so, nicely put, you should probably target M* starting now.
15:24:20 <cdent> that seems fair, we have a lot of testing that needs to happen
15:25:02 <gordc> cdent: i'm going to assume that was your topic
15:25:12 <cdent> oh yeah of course
15:25:43 <gordc> i'll get back to you.
15:25:44 <gordc> next!
15:25:54 <gordc> #topic recurring: ceilometerclient release?
15:26:07 <gordc> i just did one, it broke stuff, it got fixed... so next
15:26:21 <gordc> #topic recurring: Gnocchi status
15:26:51 <gordc> jd__: cdent: any fancy updates?
15:27:00 <jd__> 1.1.0 is out!
15:27:08 <gordc> is that influx driver ready? passes my local tests... but i haven't tried in devstack
15:27:18 <ildikov> jd__: \o/ :)
15:27:19 <gordc> jd__: \o/
15:27:20 <cdent> still needs devstack adjustments to make go
15:27:21 <jd__> http://lists.openstack.org/pipermail/openstack-announce/2015-August/000522.html
15:27:30 <jd__> gordc: works with tox locally
15:27:41 <jd__> we don't really have devstack gating so whatever
15:28:14 <gordc> cool cool. i'll wait for those adjustments.
15:28:35 <nadya> I'm setting up real env for gnocchi influx testing, but it will take 1-2 days
15:28:39 <gordc> did we ask infra whether we could gate? seeing as it's not packaged.
15:28:46 <eglynn> jd__: is the gnocchi versioning following strict semver?
15:28:51 <jd__> gordc: no because we know the answer
15:28:55 <gordc> nadya: nice
15:29:02 <gordc> jd__: word
15:29:03 <jd__> eglynn: yes sir (say I, shaking in fear)
15:29:04 <eglynn> i.e. 1.1.0 only adds backward compatible features over 1.0.0
15:29:10 <eglynn> LOL :)
15:29:19 <jd__> eglynn: yes that's it
15:29:26 <jd__> we didn't break anything (that we noticed)
15:29:30 <eglynn> coolness, thanks for the conformation
15:29:37 <eglynn> *confirmation
15:29:41 <jd__> and 1.1.1 will be bugfix if we ever need
15:29:51 <eglynn> nice
15:31:02 <gordc> next?
15:31:20 <gordc> next
15:31:23 <gordc> #topic https://review.openstack.org/212584
15:31:26 <gordc> arf
15:31:36 <gordc> #topic moar tests please!
15:31:44 <cdent> this is mine
15:31:44 <gordc> i need to learn how to copy/paste..
15:32:29 <cdent> it's just a gentle friendly reminder that in the face of all our awesome improvements (declarative meters, making aodh, having a gnocchi, removing pipeline from pollsters, functional test split out) we've disrupted our test coverage a _lot_
15:32:47 <cdent> and pretty much anytime I look into a log file in devstack I find a new warning or error
15:33:01 <cdent> and I'm trying to be dilligent about recording the bugs and fixing etc
15:33:12 <gordc> cdent: polling logs or all?
15:33:27 <cdent> but what it really means is that we need more tests in all three of the styles: unit, functional, integration
15:33:42 <cdent> gordc the most recent was in the alarm-evaluator in aodh not being able to talk to keystone properly
15:33:58 <cdent> that turned out to be a devstack/plugin.sh issue
15:34:08 <cdent> but my point is that we cannot and should not ever trust anything
15:34:11 <llu-laptop> cdent: is there a list of bugs?
15:34:24 <cdent> only the usual place llu-laptop: on launchpad
15:34:31 <gordc> cdent: good point
15:34:38 <llu-laptop> we're having a China hackathon next week, and I hope we could help to fix some bugs
15:34:39 <gordc> re: be paranoid
15:34:43 <cdent> but what I'm trying to say is that we have _created_ bugs, recently, and we dont' know about them yet
15:35:17 <cdent> llu-laptop: if you want to find the bugs, the think to do seems to be run devstack, try stuff, and look in the log files for things that surprise you, especially when using all three of aodh, ceilometer, and gnocchi
15:35:21 <cdent> it's quite fun
15:35:21 <ildikov> cdent: do you have plans on how we could track the test coverage?
15:35:24 <cdent> but also quite...alarming
15:35:39 <cdent> ildikov: not off the top of my head, no
15:35:39 <gordc> get out
15:35:46 * cdent leaves
15:36:14 <gordc> we accept no puns here.
15:36:26 <ildikov> cdent: but I guess it would be beneficial to track it somehow
15:36:33 <cdent> not even the unintentional conformation one from eglynn earlier, that one was masterful
15:36:37 <gordc> should we dedicate some time pre-freeze to testing?
15:37:09 <gordc> cdent: it went over my head... so ... it's acceptable.
15:37:22 <cdent> ildikov: yes, I agree, I just don't know what tools are really available to us. I tend to think that the easiest thing to do is just run the tests with manual coverage every now and again, and pick something to make better
15:37:58 <ildikov> we could use the RC period too for adding new tests I guess
15:38:01 <cdent> anyway, I don't wish to belabor the point, just make sure it was in everybody's brains
15:38:04 <ildikov> or is that cheating?
15:38:14 <cdent> ildikov: I think it is better to start now
15:38:18 <cdent> because the point isn't to add tests
15:38:20 <cdent> the point is to fix bugs
15:38:28 <cdent> s/fix/find and fix/
15:38:31 <cdent> and we need to do that sooner than later
15:38:46 <ildikov> cdent: what if at least run the tox -ecoverage and publish the result somewhere?
15:39:50 <ildikov> cdent: I didn't mean to start it at that time, but more to continue, when we get there, at leats we will be able to focus on it a bit
15:39:53 * cdent defers to gordc's better judgement on such things
15:40:08 * gordc doesn't really trust coverage
15:40:30 <gordc> i have time to start looking at other non-spec related stuff so i can look at testing
15:40:37 <ildikov> gordc: what do you trust?
15:40:49 * cdent doesn't trust coverage either
15:40:51 <gordc> tbh, most of the gaps we've seen came from when others have been implementing new things
15:41:36 <gordc> ildikov: nothing.
15:41:48 * cdent trusts canadians
15:41:50 <ildikov> we need more functional tests for sure
15:42:00 <ildikov> cdent: LOL :)
15:42:01 <gordc> cdent: you shouldn't, we talk a lot of trash.
15:42:20 <gordc> ildikov: agreed re: functoinal tests
15:43:05 <gordc> ildikov: i'll try runing coverage to see if anything stands out... but i don't really have a target/start point aside from that in regards to what i will fix
15:43:36 <gordc> i'm sort of hoping i'll stumble into something.
15:43:57 <ildikov> I just asked about it to have something to start with
15:44:23 <gordc> llu-laptop: i tried to clean up some bugs recently... they should (mostly) be accurate
15:44:53 <gordc> llu-laptop: you guys can take a look at it at hackathon.
15:45:04 <llu-laptop> gordc: thx very much
15:45:05 <ildikov> it could be used to see if some new features cause significant decrease then we can ask for more tests there for instance
15:46:30 <gordc> ildikov: fair enough... with so many lines of code, it's usually difficult for a patch to make that much impact on coverage
15:46:37 <gordc> even if it's all untested
15:47:12 <gordc> llu-laptop: i should mention, there's a lot of them "in progress", but it seems they aren't really
15:47:46 <ildikov> gordc: yeap, you're prolly right
15:48:33 <gordc> so i guess if you aren't working on spec or bug, start testing and writing tests...
15:48:40 <gordc> jasonamyers: you like tests right!
15:48:45 <jasonamyers> I do
15:48:53 <gordc> :)
15:49:00 <jasonamyers> I'm traveling next few weeks though
15:49:07 <gordc> well played... well played
15:49:36 <gordc> any other items on testing?
15:49:37 <ildikov> jasonamyers: good excuse :)
15:49:53 <jasonamyers> ha
15:50:12 <gordc> #topic Open discussion
15:50:30 <jasonamyers> SO today if very close to our drop dead for M* midcycle
15:50:47 <jasonamyers> I've not seen very much input from anyone except cdent and ildikov
15:50:56 <jasonamyers> is no one else interested or able to come?
15:51:15 <jasonamyers> I'm totally down to just throw a midcycle with the 3 of us, but I was hoping for me
15:51:19 <jasonamyers> s/me/more
15:51:30 <gordc> jasonamyers: doodle link again?
15:51:53 <gordc> i'm just going to put a rough date i may be free... but most likely will be 50-50
15:52:00 <jasonamyers> http://doodle.com/hmukqwzvq7b54cef
15:52:11 <jasonamyers> should we move it up a week?
15:52:15 <jasonamyers> does that help anyone if we do?
15:52:24 <jasonamyers> to be the last of november
15:53:06 <gordc> jasonamyers: possibly...
15:53:07 * cdent has tday obligations
15:53:17 <eglynn> last of nov is Thanksgiving, no?
15:53:22 <jasonamyers> yes
15:53:30 <jasonamyers> I never can remember who all celebrates that
15:53:44 <eglynn> a-ha, yeah, wot cdent said
15:53:47 <gordc> thanksgiving is in october :P
15:53:59 <cdent> i wish you were right gordc
15:54:00 <jasonamyers> that sounds very canadian :P
15:54:06 <ildikov> what is thanksgiving? ;)
15:54:08 <gordc> *shrugs*
15:54:22 <gordc> did we rule out early January?
15:54:31 <gordc> everyone be too groggy from new years?
15:54:32 <jasonamyers> I didn't
15:54:43 <jasonamyers> and I'm up for spending new years with you all :P
15:54:52 <gordc> jasonamyers: not that early
15:55:00 <jasonamyers> there was some concern it would be too late in the cycle at one point
15:55:20 <cdent> based on when everyone did there's this cycle I think that may have been a bit of a red herring
15:55:25 <jasonamyers> haha
15:55:28 <cdent> theirs
15:55:41 <jasonamyers> I guess what I'm saying is, ildikov needs to book tickets
15:55:52 <jasonamyers> I'm open from Last Nov - JAN
15:56:00 <jasonamyers> let's make this happy
15:56:03 <jasonamyers> s/happy/happen
15:56:06 <cdent> happy too
15:56:50 <gordc> jasonamyers: what's are cutoff date for deciding? soon? or week before summit?
15:57:06 <ildikov> jasonamyers: thanks, I managed to move the tickets yesterday, so now I have time until october
15:57:15 <jasonamyers> oh okay
15:57:26 <cdent> that gives a bit of breathing room
15:57:33 <jasonamyers> I need roughly 2 months of planning time
15:57:51 <jasonamyers> from myside because I'll be handling setting up stuff that is half a day away
15:58:04 <jasonamyers> so they'll be delays in organizing the facilities and food
15:58:18 <gordc> ok. so end of month...
15:58:29 <gordc> please comment folks. or just put your name and say none
15:58:45 <jasonamyers> when is summit again?
15:58:54 <ildikov> I guess regarding holidays and Christmas we might iam for January
15:59:04 <gordc> if not, we'll aim for an americas tz virtual meetup!
15:59:05 <ildikov> jasonamyers: October 27-30
15:59:27 <gordc> https://wiki.openstack.org/wiki/Liberty_Release_Schedule
15:59:33 <ildikov> gordc: works for me
15:59:41 <cdent> I'd be able to manage january much more easily from a sanity standpoint
15:59:54 <jasonamyers> okay so I'll edit the pool
16:00:05 <jasonamyers> and realign the budget stuff
16:00:10 <gordc> jan is better for me too but we'll probably have to re-evaluatate what we discuss
16:00:13 <gordc> times up
16:00:21 <gordc> let's switch channels
16:00:24 <gordc> thaakns
16:00:31 <cdent> gordc is danish
16:00:37 <gordc> #endmeeting