15:00:30 <witek> #startmeeting monasca
15:00:31 <openstack> Meeting started Wed Nov 29 15:00:30 2017 UTC and is due to finish in 60 minutes.  The chair is witek. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:32 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:00:35 <openstack> The meeting name has been set to 'monasca'
15:00:36 <sc> yo
15:00:37 <jgr> Hello
15:00:43 <fouadben> hi
15:00:43 <witek> hello
15:00:45 <cbellucci> Hello
15:00:51 <tobiajo> hi
15:00:51 <joadavis> hiya
15:01:13 <witek> just two items on the agenda today
15:01:21 <witek> feel free to add
15:01:50 <witek> #topic monasca-agent detection
15:01:56 <Neptu> hej
15:02:21 <witek> hi Neptu
15:02:42 <witek> is it your topic?
15:02:56 <Neptu> no not really
15:03:07 <witek> Tobias reported the bug in detection plugin
15:03:19 <Neptu> im working now more into k8s
15:03:20 <tobiajo> it was me that wanted to address the fact that monasca-setup often crashes when it uses oslo_config
15:03:34 <witek> trying to parse configuration using oslo.config
15:03:40 <tobiajo> link to reported issue: https://storyboard.openstack.org/#!/story/2001303
15:03:59 <witek> I had a look at this before the meeting
15:04:39 <witek> it looks to me, that parsing the config file is imlemented not robust enough, as you noted
15:04:49 <tobiajo> what did you find. anyone here that was involved implementing the usage of oslo.config?
15:05:30 <witek> the only two build-in options for oslo.config are  config_file and config_dir
15:06:01 <witek> only these should be accepted from the cmdline
15:06:19 <witek> so the bug is in monasca_setup.detection.utils.load_oslo_configuration
15:06:29 <tobiajo> what solution do you see?
15:07:31 <witek> change the code around this line https://github.com/openstack/monasca-agent/blob/stable/pike/monasca_setup/detection/utils.py#L186
15:07:51 <witek> to accept only buil-in options
15:07:56 <witek> https://docs.openstack.org/oslo.config/latest/reference/builtins.html
15:08:14 <witek> what do you think?
15:09:33 <witek> we can continue the discussion in the story
15:09:47 <tobiajo> if only config_file and config_dir should be accepted, it easy to patch. but what was the reason behind impl. usage of oslo.config here?
15:10:30 <tobiajo> i don't see the reason behind this added complexity
15:10:43 <witek> to have generic way of getting information from configuration files
15:11:01 <witek> every deployment can have them in different locations
15:12:51 <tobiajo> makes sense
15:13:50 <tobiajo> if non one else have any input it is ok for me to continue on to the next topic
15:14:21 <witek> btw. only two plugins use it at the moment, as far as I can see
15:14:45 <witek> and there is an epic story for this:
15:14:50 <witek> https://storyboard.openstack.org/#!/story/2000999
15:15:40 <witek> #topic PTG in Dublin
15:16:15 <witek> OpenStack Foundation has officially announced the next Project Teams Gathering
15:16:33 <witek> it will be held in Dublin,  February 26 - March 2, 2018
15:17:02 <witek> I've been asked if Monasca will attend the Gathering
15:17:33 <sc> I'll attend ;-)
15:17:54 <jgu> I am interested but will see if there is travel budget
15:18:01 <witek> yes, you're half Irish :)
15:18:36 <sc> witek: hahaha
15:18:51 <witek> I wanted to generally ask, if you think it's a good idea to meet in person
15:19:03 <witek> or you prefer a video conference
15:19:13 <Yavor_OP5> I think it would be beneficial
15:19:50 <witek> for people having budget problem, there is Travel Support Program
15:20:08 <witek> the details are described on the website
15:20:24 <witek> https://www.openstack.org/ptg/
15:20:44 <sc> witek: I prefer meeting in person and this time I hope to get funds for the short travel
15:21:43 <witek> yes, I also prefer meeting in person, but having representation of all parties is also important
15:22:33 <witek> so, I have set up again a Google Form to collect your opinions
15:22:59 <sc> good
15:23:13 <jgu> witek: thanks for the link. It's int'l travel for me so we'll see :-). I prefer to meet in person since it's quite a while. But I imagine we may alway need a conference line for those who couldn't travel.
15:23:49 <witek> yes, that's an option too
15:24:09 <witek> although it's not the same
15:25:07 <jgu> agreed :-)
15:25:09 <witek> I have to give the answer if Monasca is attending next week
15:26:17 <witek> so please fill in the form, and we'll update next week
15:27:04 <witek> #topic Cassandra update
15:27:45 <jgu> that's me. Just wanted to give a quick update that cassadra now has passed the new zuul gates with cassandra as tsdb.
15:27:58 <witek> great news!
15:28:05 <sc> WOW
15:28:16 <nseyvet> congrats!
15:28:24 <witek> 'Depends-On:' tags seem to work :)
15:28:24 <joadavis> +1
15:28:34 <jgu> yeah, thx all for the help :-)
15:29:01 <witek> I think we can finally get someone for review this week
15:29:52 <sc> looks strange but I'm quite free next days so I can go to my backlog of reviews
15:30:01 <jgu> that'd be great. The python persister impl now also has the smart batching feature that has potential to get similar performance as Java. We will test that out once Java performance testing is finished (only one test environment)
15:30:05 <witek> sc: would be great
15:30:21 <jgu> sc and witek: thanks in advance!
15:30:41 <witek> are you updating the performance results somewhere?
15:31:15 <tobiajo> congrats :)! good work
15:31:18 <jgu> where would be a good place to put the ifnal performanc enumber?
15:31:33 <jgu> tobiajo: thanks
15:32:17 <witek> you had that etherpad, I think that's OK for now
15:32:49 <witek> https://etherpad.openstack.org/p/cassandra_monasca
15:33:00 <jgu> sounds good. We'll refresh that etherpad and reposted the link in the IRC once it's done
15:33:26 <witek> thanks
15:33:33 <tobiajo> a little off topic: would performance have it place on docs.openstack.org/monasca-api ?
15:34:29 <witek> could be, but there are more basic things missing there right now :(
15:35:19 <witek> like installation, configuration, usage
15:36:33 <witek> do we have anything else for today?
15:37:00 <jgu> should we have a wiki for performance and tuning?
15:37:44 <witek> yes, wiki would probably be better than d.o.o
15:37:50 <tobiajo> does other OS projects have that?
15:38:07 <witek> tobiajo: I don't know, have to check
15:38:15 <jgu> what is d.o.o?
15:38:23 <witek> docs.openstack.org
15:38:32 <jgu> got it
15:38:43 <witek> btw. the wiki page needs a clean-up
15:39:17 <witek> and also, do we want to maintain both?
15:39:51 <sc> sorry, I don't understand your question?
15:40:13 <witek> do we want to maintain wiki and docs.openstack.org?
15:40:15 <tobiajo> d.o.o and wiki?
15:40:34 <jgu> I don't find Monasca on the https://docs.openstack.org/pike/projects.html
15:40:39 <witek> or should we rather deprecate wiki, and use d.o.o only
15:40:57 <sc> witek: I think that users are more likly going to docs.o.o
15:41:02 <tobiajo> i don't see any reason to have both
15:41:07 <witek> jgu: I can add it, there is just not much info there till now
15:41:30 <jgu> I agree with sc. d.o.o is more end user facing
15:42:11 <sc> wiki is for tips and tricks, user provided documentation (that deployment we forgot)...
15:43:35 <witek> I also would prefer to have d.o.o in good shape and remove most of information from wiki
15:45:23 <witek> Is there anyone interested on working on documentation?
15:45:45 <witek> I could start with defining the tasks
15:47:40 <witek> so coming back to the origin question, yes we could probably add performance section as well
15:48:05 <jgu> sure. I am generally no good/help at documentation but I'll check out the task list you are going to create and see if I can help some.
15:48:08 <sc> writing docs is a nighmare, we have to have docs but ... could we autogenerated from docstrigns? ;-)
15:48:54 <witek> for api-ref possibly yes
15:49:57 <witek> there is also that change for python client:
15:50:00 <witek> https://review.openstack.org/#/c/511341/
15:50:33 <witek> click on build-openstack-sphinx-docs job output to see the rendered docs
15:52:42 <witek> if there is nothing else, I think we can close the meeting
15:53:21 <witek> please think about PTG and the google form
15:53:30 <joadavis> I'll add a quick status on monasca-ceilometer.  After the Zuul v3 update we found the gates failed even on some simple changes.  Disabled the requirements-check gate (was tripping on a git+https reference for pulling in ceilometer) and now I'm working on bringing the code up to date with Ceilometer Pike. The intent is to get in sync before submitting the publisher to Ceilometer, an action item that came out of the Monasca
15:53:30 <joadavis> eens mid-cycle.
15:53:32 <witek> bye, see you next week
15:54:08 <witek> joadavis: ok, thanks
15:54:59 <witek> so, are zuul jobs working now?
15:55:08 <witek> for monasca-ceilometer?
15:56:10 <joadavis> There is a pending commit that needs a +2, but that gets around the issue. https://review.openstack.org/#/c/518414/  And a change was put in the upstream project config that made it possible.
15:57:00 <witek> ok, adding myself to review
15:57:34 <joadavis> Thanks. :)
15:57:58 <witek> thank you everyone
15:58:06 <witek> bye
15:58:12 <haruki> thank you
15:58:15 <joadavis> bye
15:58:18 <cbellucci> bye
15:58:22 <tobiajo> bye
15:58:27 <witek> #endmeeting