15:00:03 <rhochmuth> #startmeeting monasca
15:00:04 <openstack> Meeting started Wed Feb 10 15:00:03 2016 UTC and is due to finish in 60 minutes.  The chair is rhochmuth. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:05 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:00:07 <openstack> The meeting name has been set to 'monasca'
15:00:11 <rhochmuth> o/
15:00:16 <tgraichen> hi
15:00:16 <rbak> o/
15:00:19 <ho_away> hi
15:00:21 <pradipm> hi
15:00:31 <tgraichen> no video today? :)
15:00:33 <witek> hi
15:00:36 <bklei> o/
15:00:45 <rhochmuth> no video today
15:01:11 <rhochmuth> well, no one has post'd any agenda items/topics
15:01:21 <rhochmuth> seeing as we had a mid-cycle last week that is understandable
15:01:37 <rhochmuth> i do have a few topics to cover
15:01:49 <rhochmuth> hopefully i remember them all
15:02:31 <rhochmuth> bklei: we had to back out some code you checked in back in October
15:02:41 <rhochmuth> did you see that?
15:03:03 <bklei> hmm, i missed that -- which code?
15:03:25 <rhochmuth> https://review.openstack.org/#/c/278046/
15:03:31 <rhochmuth> Revert "Avoid multiple inner joins if dimensions specified"
15:03:43 <bklei> you saw problems with that?
15:03:49 <rhochmuth> There were some bugs with pagination and specifying limits
15:04:04 <rhochmuth> we had some bugs and traced it to that commit
15:04:12 <rhochmuth> we didn't understand how to resolve either
15:04:18 <rhochmuth> so, reverted it
15:04:57 <bklei> hmm, that's a bummer, that change made a significant perf improvement, would like more info on the bugs, i can take a look and fix
15:05:01 <rhochmuth> If you have ryan bradt's email you can get more details
15:05:06 <rhochmuth> unless he is lurking
15:05:11 <bklei> ok, will follow up with him
15:05:26 <rhochmuth> i don't see him around yet
15:05:59 <rhochmuth> ok, my next topic is we would like to extend some of the filters that are allowed
15:06:22 <rhochmuth> basically when filtering on alarms or alarm definitions we would like to specifiy a '|' operator
15:06:34 <rhochmuth> on the alarm state and severity
15:07:01 <rhochmuth> this would allow you to get all the alarms in the ALARM|UNDETERMINED state
15:07:18 <rhochmuth> or with a severity of "LOW | MEDIUM"
15:07:37 <rhochmuth> it is a small change and it should work
15:07:50 <bklei> so this is a filter applied when doing a GET?
15:07:59 <rhochmuth> Correct
15:08:10 <rhochmuth> GET alarms
15:08:12 <bklei> ahh, sounds like a good change
15:08:16 <rhochmuth> GET alarm-definitions
15:08:29 <bklei> cool
15:09:07 <rhochmuth> rbrandt has added a number of filters and order_by parameters recently
15:09:17 <rhochmuth> most are still in reveiw waiting to get merged
15:09:24 <rhochmuth> so this is the last one that we know about
15:09:34 <rhochmuth> other than the possiblity of adding a dimensions resource that we talked about
15:10:06 <witek> could you please add someone from FJ as reviewer for all ORM related stuff
15:10:13 <rhochmuth> the main reason is to allow user-interfaces, like Horizon, to be able to filter and paginate through list of alarms and definitions without client side processing
15:10:58 <rhochmuth> witek: are you referring to michal's review
15:11:15 <witek> no, extensions for filtering
15:11:26 <rhochmuth> sure
15:11:46 <witek> thanks
15:11:47 <rhochmuth> yes, this will add some new parameters to the repositories interfaces
15:12:03 <rhochmuth> so fj needs to be looking at this
15:12:17 <rhochmuth> it looks like there will be some merge conflicts as a result too
15:12:26 <rhochmuth> that will need to be rebased
15:12:28 <rhochmuth> what fun
15:13:17 <rhochmuth> so, i'm guessing that folks are ok with that change, but the reveiw will be up soon so folks can take a look and comment on it
15:13:44 <rhochmuth> so, another topic is tornado in the log api
15:14:01 <rhochmuth> i saw the response to my analysis yesterday
15:14:19 <rhochmuth> i'll dive into my code further
15:14:29 <rhochmuth> it could be that i've got a problem
15:15:00 <witek> ok, thanks for info
15:16:04 <rhochmuth> This is the review i was referring too, https://review.openstack.org/#/c/277818/
15:17:07 <rhochmuth> OK, next topic is the ORM work
15:17:23 <rhochmuth> i had some comments
15:17:46 <rhochmuth> i was wondering if we should explore moving to the sqlalchemy implementation
15:18:17 <rhochmuth> here is the review, https://review.openstack.org/#/c/266922/
15:18:49 <witek> quite a few comments, i'm not up-to-date, must admit
15:19:24 <rhochmuth> so, i just wanted to discuss the possiblity of moving to the sqlalchemy implementation in the Python API
15:20:20 <rhochmuth> i haven't socialized that idea internally, but i will discuss sometime today or tomorrow what the implications/thoughts are
15:20:21 <witek> sure, for us it would be great
15:21:14 <rhochmuth> ok, i thought so
15:22:15 <rhochmuth> so, that was all that i had
15:22:54 <witek> could you approve the blueprint, please? https://blueprints.launchpad.net/monasca/+spec/alarmsonlogs
15:23:53 <rhochmuth> ok, i've approved
15:24:07 <witek> thanks!
15:24:27 <rhochmuth> it seems like this won't make it into the mitaka release cycle
15:25:00 <rhochmuth> I did want to mention that voting is up for OpenStack
15:26:23 <bklei> yeah -- like 9 talks that mention monasca!
15:26:27 <rhochmuth> Here is the search form
15:26:29 <rhochmuth> https://www.openstack.org/summit/austin-2016/vote-for-speakers/SearchForm
15:26:47 <rhochmuth> that is a good number
15:26:56 <witek> nice
15:27:02 <rhochmuth> including a couple for some folks that I didn't know were working in this area
15:27:21 <rhochmuth> so, please take a look at all the submissions
15:28:38 <rhochmuth> Fabiog has got the python-monascaclient going in the global requirements again at
15:28:39 <rhochmuth> https://review.openstack.org/#/c/251674/
15:28:53 <rhochmuth> it looks like everything is green, except for a non-voting gate
15:29:00 <rhochmuth> hopefully this gets through this time
15:29:52 <rhochmuth> Also, my changes in governance look like they are progressing at, https://review.openstack.org/#/c/274884/
15:31:00 <rhochmuth> so, any other topics, status updates, questions, ...
15:31:15 <pradipm> hello
15:31:22 <rhochmuth> hi pradipm
15:31:28 <pradipm> if there is no topic, may I allowed to ask 2/3 questions?
15:31:37 <rhochmuth> sure go ahead
15:31:47 <pradipm> thanks!
15:32:14 <rbak> I do have a grafana update also, if anyone is interested
15:32:18 <pradipm> I am not able to list alarm-defn from cli
15:32:39 <rhochmuth> ok, that should work
15:32:41 <pradipm> getting service-unavailable
15:32:53 <rhochmuth> what environment are you using
15:33:03 <rhochmuth> devstack, monasca-vagrant, ...
15:33:11 <pradipm> however notification-list, alarm-defn creation etc. is working
15:33:14 <pradipm> devstack
15:33:47 <rhochmuth> well, that doesn't make sense
15:33:51 <rhochmuth> it is possible there is a bug
15:34:07 <rhochmuth> i don't have an env up right now
15:34:14 <pradipm> is there a way I can paste the cli output here?
15:34:23 <rhochmuth> sure, paste it
15:34:29 <rhochmuth> also, as we are getting off topic
15:34:33 <pradipm> stack@ubuntu:~/devstack$ monasca --os-username admin --os-password secretadmin --os-project-name admin alarm-definition-list
15:34:34 <pradipm> ERROR (exc:65) exception: {
15:34:34 <pradipm> "title": "Service unavailable",
15:34:34 <pradipm> "description": ""
15:34:34 <pradipm> }
15:34:34 <pradipm> HTTPException code=500 message={
15:34:35 <pradipm> "title": "Service unavailable",
15:34:35 <pradipm> "description": ""
15:34:36 <pradipm> }
15:34:58 <rhochmuth> maybe just contact me in IRC directly or in the openstack-monasca room after the meeting
15:35:09 <pradipm> sure. thanks a ton.
15:35:18 <pradipm> another question  - may I ask pls?
15:35:24 <rhochmuth> sure
15:35:44 <pradipm> is it possible to launch Graphanna from Horizon plugin of Monasca?
15:35:55 <rhochmuth> yes
15:35:58 <pradipm> I can see the monitoring tab .. view all info
15:36:22 <rhochmuth> there is a button in the upper left and grafana can be launched from that
15:36:24 <pradipm> but when graffana is launched .. seeing an error : Request Method:         GET                       Request URL:         http://10.74.150.152:8000/grafana/index.html
15:36:49 <rhochmuth> so, let's take that off-line too
15:37:00 <pradipm> I am in horizon plugin as per : https://github.com/openstack/monasca-ui/
15:37:05 <pradipm> Sure.
15:37:14 <bklei> oh -- we hit that, ping me pradipm
15:37:24 <bklei> basically need an apache alias
15:37:24 <pradipm> last question. extremely sorry for so much asks.
15:37:25 <rhochmuth> ohh, i'm not sure the horizon plugin is integrated in the devstack environment
15:37:41 <pradipm> sure bklei.
15:38:15 <pradipm> If I run monasca in scale-out mode ... what all components/services are shared?
15:38:34 <rhochmuth> what do you mean by shared?
15:38:40 <jobrs> @rbak: highly interested :-)
15:38:47 <rhochmuth> "clustered?"
15:38:55 <pradipm> I mean, say the influxDB is common for all the deployment
15:39:45 <rhochmuth> zookeeper, kafka, influxdb, storm and mysql are all shared
15:39:57 <rhochmuth> and clustered too
15:40:20 <pradipm> Ok. thanks a lot for all the answers.
15:40:33 <rhochmuth> ok, see you over in #openstack-monasca
15:40:41 <rhochmuth> rbak: update on grafana?
15:40:48 <pradipm> Sure, thanks rhochmuth.
15:41:05 <rbak> So the patches are ready and pull requests are up.
15:41:13 <rhochmuth> as in completely done
15:41:22 <rbak> Yes
15:41:29 <rhochmuth> any feedback on your pull requests?
15:41:32 <rbak> The monasca datasource is not going to get merged.
15:41:40 <rbak> They're moving away from that model
15:41:48 <rbak> Even the plugins repo will be deprecated
15:42:00 <rhochmuth> so, there is grafana-uplugins,
15:42:17 <rbak> grafana-plugins is going to be deprecated
15:42:20 <rhochmuth> that is where the datasource is located
15:42:30 <rhochmuth> so, will you just have your own repo
15:42:47 <rhochmuth> and what about the Keystone integration?
15:42:49 <tgraichen> but we can provide it as third party plugin like i think itwill be the case for any new one - right?
15:43:07 <rbak> Correct. Plugins will maintained outside grafana, and they're creating a new website to make finding plugins easier.
15:43:40 <rhochmuth> so, do you want to create a monasca-grafana repo in the openstack organization
15:43:42 <rbak> I'm also told that plugins will be more stable in the future because of this and not require constant updates.
15:44:07 <rbak> rhochmuth: That would probably be a good thing at some point
15:44:26 <rhochmuth> yeah, i think so
15:44:31 <rbak> The keystone auth is trickier.
15:44:38 <rbak> The link to the PR is here: The monasca datasource is not going to
15:44:46 <rbak> Sorry, here: https://github.com/grafana/grafana/pull/3963
15:45:18 <rbak> But there is some resistance to merging a "non standard auth system"
15:45:26 <rhochmuth> i see
15:46:29 <rbak> I've been asked to look into the auth proxy grafana already provides, and I've agreed, but I know it won't work as is and I'm fairly sure it would require huge changes to make it work with the keystone auth
15:46:30 <rhochmuth> well, sounds like he is open to it
15:46:48 <rbak> Sure, I think it just needs some support
15:47:09 <rhochmuth> well, i think it is up to everyone here
15:47:16 <rhochmuth> can we vote for pull requests
15:47:42 <rbak> I'm not sure
15:47:49 <rbak> It may just need comments
15:47:51 <rhochmuth> also, i thought you said this would be applicable to Ceilometer too
15:47:59 <rbak> It will be
15:48:16 <rbak> It could be used by any openstack service
15:48:51 <rhochmuth> i'll contact ceilometer as well
15:49:02 <rhochmuth> not sure how there existing plugin does auth
15:49:05 <rhochmuth> do you know
15:49:14 <rbak> They have two options currently
15:49:21 <rbak> One is putting in a token
15:49:27 <rbak> Like we had before
15:49:47 <rbak> The other is to put a username and password in the datasource and use that to get tokens.
15:51:08 <rhochmuth> so, i guess you'll update torkel at some point
15:51:32 <rhochmuth> in the mean time i think we should leave comments asking for the merge
15:51:41 <rbak> Sounds good
15:51:51 <rhochmuth> i'll leave some comments later today
15:52:42 <rhochmuth> he specifically mentions +1 votes
15:53:25 <rbak> Looking at other pull requests, it looks like people vote in the comments.
15:54:25 <rhochmuth> thanks rbak
15:54:44 <rhochmuth> so, we are getting to the end
15:54:50 <rhochmuth> any other topics?
15:55:36 <rhochmuth> going once
15:55:42 <rhochmuth> …
15:56:10 <rhochmuth> ok, thanks everyone
15:56:18 <rhochmuth> see you next week or in irc
15:56:22 <witek> thank you Roland!
15:56:48 <rhochmuth> #endmeeting monasca