15:13:08 <jd__> #startmeeting ceilometer
15:13:08 <openstack> Meeting started Thu Aug 22 15:13:08 2013 UTC and is due to finish in 60 minutes.  The chair is jd__. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:13:09 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:13:12 <openstack> The meeting name has been set to 'ceilometer'
15:13:14 <jd__> awesome thansk jeblair
15:13:21 <jeblair> np
15:13:21 <jd__> #topic Review Havana-3 milestone
15:13:24 <terriyu> o/
15:13:32 <thomasm> o/ again =]
15:13:33 <jd__> #link https://launchpad.net/ceilometer/+milestone/havana-3
15:13:33 <dragondm> Whee :>
15:13:35 <sileht> o/
15:13:36 <llu-laptop> o/
15:13:51 <jd__> so there's only 2 weeks left to propose your patches
15:14:10 <jd__> and seeing how long and hard it is to get patches merged, you shouldn't wait for last minute clearly
15:14:35 <jd__> we're in good shape I think now, eglynn's coming with some patches as soon as it comes back from vacation
15:14:45 <jd__> terriyu made terrible progress on her blueprint
15:14:51 <gordc> :)
15:14:52 <terriyu> what?
15:15:02 <jd__> terrible as in awesome
15:15:09 <terriyu> phew
15:15:11 <jd__> (not sure that was clear? :)
15:15:17 <jd__> sorry, french-ism
15:15:20 <terriyu> I hope that's not a joke on my name :)
15:15:34 <jd__> haha, no I didn't notice
15:15:52 <jd__> well, and the other blueprint are either low or needing code reviews
15:16:04 <terriyu> yes, I'm working on the group by API blueprint
15:16:11 <jd__> we have a long patch queue right now, it'd be good to have more review btw
15:16:16 <terriyu> #link https://blueprints.launchpad.net/ceilometer/+spec/api-group-by
15:16:23 <terriyu> I'll try to keep the blueprint up-to-date
15:16:37 <terriyu> so far I wrote the storage tests and jd did the SQL Alchemy implementation
15:16:40 <jd__> and terriyu's patches are a good start for review indeed :)
15:16:44 <terriyu> I just did the MongoDB implementation
15:16:45 <maksimov> o/
15:16:50 <gordc> will start reviewing again now that i got my patch passing tests
15:17:20 <terriyu> so all that's left is the API implementation
15:17:30 <terriyu> might need your guys help on understanding the API
15:18:51 <llu-laptop> I'm supposed to have more time for reviewing next week, since not many nova guys are reviewing our patches there. :(
15:19:02 <jd__> no problem terriyu
15:19:18 <jd__> llu-laptop: cool, somehow
15:19:47 <gordc> llu-laptop: our review page is pretty good once you look at nova's
15:20:00 <jd__> yeah
15:20:04 * dragondm will contribute +1's ;->
15:20:20 <herndon> but no -1's right? :p
15:20:20 <llu-laptop> gordc: I just counted nova's reivew page, there are more than 400 pendings
15:20:34 <dragondm> heh those too, as needed.
15:20:44 <gordc> llu-laptop: another project branched off nova?
15:21:29 <dragondm> nova's sortof an amoeba, keeps dividing.
15:21:49 <dragondm> but gets bigger again afterword :>
15:22:00 * sileht participes for the first time to the openstack freeze, the zuul queue and the reviews numbers is amazing
15:22:29 <jd__> #topic Open discussion
15:23:20 <llu-laptop> jd__: about the hosts configuration in the pipeline.yaml in patch #link https://review.openstack.org/43072
15:23:36 <jd__> yes, I didn't think it through yet
15:24:27 <jd__> I think it's on the good way but I don't like it as it is, I'd like something more generic if possible
15:24:50 <jd__> and if I can't think about something, having another YAML file would be an option I guess
15:24:55 <llu-laptop> since those hosts configuration is only used by hardware agent, and is kind of independent of pollsters, i don't know how to put it into pipeline.yaml more nicely.
15:25:24 <llu-laptop> I'm ok with another yaml file, it was my first thought.
15:25:43 <sileht> llu-laptop, jd__ I prefer a other file too
15:25:48 <gordc> llu-laptop: is there a list of hardware metrics that we are targeting? i have some ppl interested in hardware metering
15:27:33 <llu-laptop> I'm doing the upstream work based on the code Toni implemented, currently there are only a few metrics there, e.g. cpu utilizations, memory, disk, net, etc.
15:28:23 <gordc> llu-laptop: ok, so whatever you have so far and we'll add stuff later if needed.
15:29:05 <jd__> llu-laptop: what we may need is a new filter in pipeline for 'resource' so you can request some meter only for a particular resources
15:29:27 <jd__> llu-laptop: that'd allow to use it to define which meter you want for which hosts
15:29:54 <jd__> llu-laptop: now, for the inspectors to use for each host, I don't know how to / if we can put that into the pipeline
15:30:15 <jd__> llu-laptop: ah we could use an URL scheme for resources indicating the inspector maybe?
15:30:51 <jd__> llu-laptop: like I want the 'cpu' meter on resources 'snmp://foobar'
15:30:56 * jd__ throws ideas
15:32:18 <llu-laptop> jd__: so you're saying adding parameters to the meter part of the pipeline?
15:33:44 <jd__> llu-laptop: 'resources' could be at the same level as 'meters'
15:35:10 <llu-laptop> jd__: sorry, I don't catch what your mean by defining 'resources', what does it used for?
15:35:51 <jd__> llu-laptop: indicating the host you want to poll for via, e.g. snmp
15:39:05 <jd__> ok I think I'll close the meeting if nobody has anything to say
15:39:09 <llu-laptop> jd__: so does this mean we coudl use only one YAML file? But this may requires change the PublisherContext.
15:39:13 <jd__> we can continue with llu-laptop on #openstack-metering
15:39:28 <jd__> llu-laptop: yes that would allow us to have one file indeed
15:40:26 <llu-laptop> jd__: ok, let's continue on ceilometer channel.
15:40:32 <jd__> #endmeeting