15:01:40 <rhochmuth> #startmeeting monasca
15:01:41 <openstack> Meeting started Wed Nov 18 15:01:40 2015 UTC and is due to finish in 60 minutes.  The chair is rhochmuth. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:01:42 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:01:44 <openstack> The meeting name has been set to 'monasca'
15:01:50 <rhochmuth> o/
15:01:56 <bklei> o/
15:01:56 <fabiog> o/
15:01:57 <tomasztrebski> o/
15:02:00 <witek> hi
15:02:01 <bmotz> o/
15:02:04 <ddieterly> o/
15:02:10 <rbak> o/
15:02:14 <s-kawabata_> o/
15:02:25 <rhochmuth> hello everyone
15:02:32 <bklei> good morning!
15:02:35 <ddieterly> \o/
15:02:38 <rhochmuth> Please add to agenda at, https://etherpad.openstack.org/p/monasca-team-meeting-agenda
15:02:43 <witek> good afternoon :)
15:02:50 <tgraichen> good afternoon
15:02:55 <tomasztrebski> more or less evening even :-)
15:02:55 <rhochmuth> gooten morgen
15:02:57 <bklei> :)
15:03:07 <s-kawabata_> good night :)
15:03:18 <tomasztrebski> 0-0
15:03:37 <rhochmuth> #topic GPL in monasca-{common, api}
15:03:45 <ddieterly> Gunter glieben glauchen globen
15:04:00 <rhochmuth> anyone got something to say
15:04:15 <bklei> besides we like def lepard?
15:04:25 <rhochmuth> The topic is the GPL MySQL library in monasca-common
15:04:30 <ddieterly> bklei: +1
15:04:30 <witek> yes, we have that mysql-python dependency
15:04:40 <witek> which brings GPL2 license
15:04:54 <rhochmuth> That library isn't an issue for HP
15:05:02 <witek> Apache2 is not compatible with GPL2
15:05:02 <rhochmuth> We reviewed with our legal
15:05:20 <witek> http://www.gnu.org/licenses/license-list.html#apache2
15:05:44 <rhochmuth> I don't really want to get into why HP Legal feels like we are covered
15:05:56 <witek> and frankly, we have very little chances that our legal department will let us go with it
15:06:03 <rhochmuth> and would rather focus on getting this resolved for the entire community
15:06:25 <rhochmuth> so, there is an alternative MIT licencsed mysql library
15:06:38 <witek> yes, pymysql
15:06:41 <rhochmuth> is it a drop in replacement for the MySQL library
15:07:24 <witek> we could also take sqlalchemy to make DB replaceble
15:07:34 <fabiog> +1
15:07:35 <rhochmuth> If it is, then I think we could modify the requirements file, do some testing to confirm no issues and move on
15:07:57 <bklei> +1 on avoiding lawyer discussions
15:08:21 <ddieterly> bklei: the party of the first part agrees with the party of the second part...
15:08:40 <rhochmuth> So, SQL alchemy is a much bigger fix
15:08:46 <rhochmuth> change
15:09:06 <witek> yes, but we would definitely be interested in it
15:09:28 <witek> without orm though
15:09:46 <rhochmuth> We could debate SQLAlchemy vs hand-written SQL statements and never get anywhere
15:10:21 <witek> what do you mean?
15:10:45 <rhochmuth> SQLAlchemy can often get in the way
15:11:17 <tomasztrebski> simply that there are as many advantages as there is disadvantages for ORM solutions
15:11:34 <rhochmuth> Yes
15:12:01 <witek> you can use sqlalchemy without orm
15:13:19 <tomasztrebski> I believe that this is something we did in monasca-notification, right ?
15:13:19 <rhochmuth> sorry, i'm not an expert on sqlalchemy
15:14:57 <witek> anyway, i think replacing mysql-python with pymysql ist worth a try
15:14:57 <tomasztrebski> ok, maybe let's not debate over that matter right now, we have more urgent issue about this GLP stuff
15:15:41 <rhochmuth> OK, so if the pymysql library is a drop in replacment, that the rest of OpenStack is using, then I would say let's convert over and see if it works
15:15:55 <tomasztrebski> there are also other things, that basically from what I learned openstack projects are more or less MIT, Apache2...BSD is less common, and if there's a GPL it is most likely LGLP which is compatible with Apache2
15:15:56 <rhochmuth> we did somethign similar with the drizzle drive for Java recently
15:16:58 <tomasztrebski> as well as the library seems to be active, as I mentioned in mailing list MySQL-Python had recent activity at github over 2 year ago
15:17:09 <tomasztrebski> in opposite to pymysql
15:17:15 <rhochmuth> 2 years is a long time
15:17:42 <rhochmuth> so, let's try converting, then we can discuss sqlalchemy another day
15:17:44 <tomasztrebski> I think that anyone'd agree that this is rather a huge gap in IT
15:17:49 <tomasztrebski> +1
15:17:57 <witek> +1
15:18:13 <rhochmuth> What is the huge gap?
15:18:29 <tomasztrebski> 2 years in development of..pretty much any application, library
15:18:41 <rhochmuth> oh yes
15:19:13 <tomasztrebski> just to close down that subject, porting to pymysql means changes in monasca-{api, thresh, common}, is there is something I forgot ?
15:19:27 <s-kawabata_> not maintained software is dead I often heard.
15:19:54 <rhochmuth> I think that is it
15:20:05 <rhochmuth> Changes to requirements are only necessary, right?
15:20:15 <witek> have to test
15:20:22 <rhochmuth> correct
15:20:39 <rhochmuth> no changes to monasca-thresh as that is java only right now
15:20:57 <rhochmuth> monasca-persister doesn't use MySQL
15:21:16 <rhochmuth> monasca-notification has to be changed
15:21:47 <rhochmuth> monasca-api, monasca-notification, monasca-common
15:21:51 <rhochmuth> i think that is the list
15:22:00 <witek> #action test pymysql replacement for mysql-python
15:22:27 <rhochmuth> i think that is relatively easy with the new Tempest tests that are in CI
15:23:02 <rhochmuth> OK, how about a new topic
15:23:07 <tomasztrebski> +1
15:23:13 <rhochmuth> that was a lot of fun though
15:23:18 <tomasztrebski> ;-)
15:23:40 <witek> +1
15:23:44 <rhochmuth> what topics from last week didn't we cover
15:23:57 <rhochmuth> that we should go into detail on
15:24:03 <bklei> none from me
15:24:03 <rhochmuth> or discuss
15:24:09 <tomasztrebski> I'd like to go with my tempest test and one question from me
15:24:23 <rhochmuth> ok
15:24:32 <rhochmuth> #topic tempest
15:24:52 <rhochmuth> tomasz, the floor is yours
15:25:03 <tomasztrebski> recently I finally managed to find a time to get back to alarm state history, basically with the fix in gerrit I proposed old tests in AlarmStateHistory are passing in 100%, plus I wrote a new one
15:25:23 <tomasztrebski> I tried to present changing the offset (i.e. paging) with changed limit, so every possible combination
15:25:33 <tomasztrebski> it also passes
15:25:53 <rhochmuth> Is this "Fix for alarm history paging"
15:26:01 <tomasztrebski> yeah, that's one
15:26:49 <tomasztrebski> using monasca-api github-master, I have failures, not only in new test but also in those that were already there
15:27:29 <ddieterly> tomasztrebski: not all tests are passing currently
15:27:34 <ddieterly> we are working on that
15:27:58 <tomasztrebski> but do you mean not all regarding alarm state history ? I concentrated only on that specific test
15:28:24 <tomasztrebski> and yesterday when I finished I repeated this this about 9 times and had 9 success
15:28:29 <rhochmuth> These changes were in the Java API, and the Java API, was close to 100%.
15:28:41 <rhochmuth> I think I was seeign only 2 tests fail.
15:28:42 <ddieterly> i'm not sure which tests are not passing, but there are a lot that don't pass with the python
15:28:55 <rhochmuth> There are 17  passes in the Python API that don't pass
15:29:07 <rhochmuth> As far as I know, only 2 fail in Java
15:29:17 <rhochmuth> So, we are trying to get that resolved ASAP
15:29:44 <rhochmuth> So, what should we do?
15:30:00 <rhochmuth> Seem liek you think the code is ready to go
15:30:22 <tomasztrebski> yes, but only for the java, honestly I did not even think of python:(
15:30:27 <rhochmuth> So, we could take a look and run the tests too
15:30:33 <tomasztrebski> I think I was too happy that I saw success :D
15:30:53 <rhochmuth> OK, we'll take a look and render a verdict
15:31:01 <tomasztrebski> please do, thx
15:31:04 <rhochmuth> trying to keep the legal theme going today
15:31:31 <bklei> :)
15:31:34 <rhochmuth> ok, next topic
15:31:40 <rhochmuth> thanks for the test tomasz
15:31:50 <rhochmuth> and the bug fix
15:31:53 <tomasztrebski> if I may, just short question only
15:31:55 <tomasztrebski> np :)
15:31:58 <rhochmuth> sure
15:32:52 <rhochmuth> schnell, schnell
15:32:59 <tomasztrebski> do you have any recipe regarding monitoring an app ? For instance I was thinking on adding some metrics to monasca log api, but clearly I have no idea how to do that ? I was a project - monasca-statsd, but I am not sure if that's a right way to go, if I'd like to use monasca-agent to collects stats from monasca-log-apo
15:33:10 <tomasztrebski> and ideally create some alarms
15:33:13 <witek> Tomasz is not German, he's Polish :)
15:33:34 <rhochmuth> Sorry, i couldn't resist :-)
15:33:43 <tomasztrebski> "szybciej, szybciej" or "ruchy ruchy" if you please :)
15:33:52 <tomasztrebski> that's the polish way...I mean polite one :)
15:34:04 <tgraichen> don't know if it fits here - i had a look at the grafana2 integration from rbak and it looked quite good to me so far - is there any rough timeline for the real keystone integration into the monasca datasource in grafana2?
15:34:18 <rhochmuth> Getting back the the question
15:34:45 <rhochmuth> i think monasca-statsd is one good way
15:35:08 <rhochmuth> it is basically a statsd library with the ability to provide dimensions
15:36:03 <tomasztrebski> mhm, but do you guys plan to use that for instance for python monasca-api ? I mean, you surely want to have some sort of metrics for main brick in whole system, right ?
15:36:33 <rhochmuth> yes
15:36:50 <rhochmuth> we already instrument the python notification engine using the statsd library
15:37:01 <tomasztrebski> and I'd like to keep monasca-log-api in that subject the same as the rest of components
15:37:13 <rhochmuth> so the goal in the python code is to use statsd for all monitoring of monitoring
15:37:40 <rhochmuth> so, does that cover that question
15:37:59 <tomasztrebski> sounds funny :), ok so statsd it is. I don't ask about an agent, it is most likely easy to collect that data later
15:38:06 <tomasztrebski> yes, it does, thank you
15:38:33 <rhochmuth> in hos we deploy the monasc agent on the systems that are running monasca too, and then use statsd for monitoring them in the case of the python components
15:38:40 <rhochmuth> so we are monitoring monitoring
15:38:54 <tomasztrebski> sounds cool
15:39:32 <rhochmuth> tgraichen rbak is the engineer doing the grafana 2 integration
15:39:47 <rhochmuth> rbak are you there?
15:39:57 <rhochmuth> any updates on grafana 2
15:40:09 <rbak> Nothing at the moment.
15:40:30 <rbak> I got pulled off to work on some other things so hopefully I can get back to grafana soon
15:40:44 <tgraichen> do you have a complete idea of how the keystone integration should be at the end or are you still in planning stages?
15:41:44 <rbak> Just planning right now.  The auth isn't pluggable so this looks like it's going to be a more complicated change.
15:42:06 <rbak> As soon as I know more I'll put the information on the blueprint
15:42:19 <tgraichen> ok - thanks a lot
15:42:30 <rhochmuth> ok, next topic
15:42:49 <rhochmuth> #topic start/end time args for metrics list change ready for review https://review.openstack.org/#/c/241626/
15:42:54 <bklei> that's me
15:43:18 <bklei> ^^^ is ready from my perspective, bmotz has looked at it, would appreciate more eyes
15:43:30 <rhochmuth> OK, we'll get to it
15:43:35 <bklei> gracias
15:43:59 <rhochmuth> look like you did java, python and tempest tests
15:44:14 <rhochmuth> so, i agree looks complete
15:44:16 <bklei> yup -- had to blow the dust off my influxdb knowledge :)
15:44:29 <bklei> thx to ddieterly for devstack!
15:44:38 <ddieterly> ur welcome!
15:44:39 <bmotz> python looked fine to me, but I did't look at the Java yet
15:44:39 <rhochmuth> that was me that did that
15:44:49 <ddieterly> uh, yea...
15:44:51 <bklei> then thx to you rhochmuth!
15:44:57 <rhochmuth> sorry, inside joke, couldn't resist
15:45:02 <bklei> great for this type of change that touches lots of stuff
15:45:05 <rhochmuth> ddieterly did it
15:45:06 <ddieterly> we need to get this review out https://review.openstack.org/#/c/244912/18
15:45:11 <bklei> oh, then i take it back
15:45:30 <rhochmuth> joes are tough in irc
15:45:36 <rhochmuth> jokes are touch in irc
15:45:38 <ddieterly> allows easier development by binding services to externally accessible ip
15:45:51 <rhochmuth> ok, i'll probably just +2
15:46:25 <bklei> lgtm
15:46:37 <rhochmuth> #topic caching
15:46:44 <rhochmuth> i started writing some code
15:46:52 <bklei> cool
15:46:57 <rhochmuth> unfortunately, i keep getting piled on with meetings
15:47:02 <bklei> :(
15:47:05 <rhochmuth> i don't have a ETA
15:47:18 <bklei> np -- let me know when there's a review to look at
15:47:29 <rhochmuth> how long can you wait on this?
15:47:45 <bklei> sheesh, you know how things are
15:47:48 <ddieterly> rhochmuth: do we have a design for that?
15:48:03 <rhochmuth> i'll review with you
15:48:21 <rhochmuth> or whoever else want to be involved
15:48:38 <rhochmuth> basically adding some hash maps to the measurements repo, similar to the persister
15:49:05 <rhochmuth> to map from (region, tenant, name, dimensions) => definition_dimensions_id
15:49:05 <bklei> are you looking at statistics calls, or statistics and measurements?
15:49:29 <rhochmuth> i'm just look at the measurements right now, but the mechansims is basiclaly the same
15:49:32 <ddieterly> is this going to be going into python and java?
15:49:48 <rhochmuth> only Java, as this is a Vertica improvment only
15:49:52 <bklei> no vertica in python, probably not needed
15:49:56 <bklei> yeah, what he said
15:50:19 <rhochmuth> #topic midcycle
15:50:28 <rhochmuth> it is not too soon
15:50:34 <rhochmuth> to start discussing
15:50:34 <bklei> cool
15:50:48 <bklei> ideas on location/hosting?
15:51:08 <bklei> twc is in office transition, probably can't host
15:51:19 <witek> when is it planed?
15:51:19 <rhochmuth> San Jose or Berlin
15:51:34 <fabiog> rhochmuth: we can host
15:51:43 <fabiog> rhochmuth: both in San Jose or Berlin
15:51:57 <rhochmuth> Yes, Cisco was going to host
15:52:12 <rhochmuth> they have offices in San Jose and Berlin
15:52:22 <bklei> fabiog maybe you put up another voting site for dates?
15:52:32 <fabiog> bklei: yes
15:52:32 <rhochmuth> and locations
15:52:36 <bklei> +1
15:52:38 <fabiog> will do that
15:52:55 <rhochmuth> just a quick check then, will folks that are in europe be able to travel to San Jose
15:53:01 <fabiog> if we do it in SJC I can try to coordinate a joint day/session with Congress
15:53:24 <rhochmuth> That would be good to add
15:53:32 <witek> have to check
15:53:50 <witek> which month?
15:54:00 <fabiog> for SJC in January
15:54:04 <tgraichen> i think for me and joachim it might be complicated to get to the us - berlin would be perfect though :)
15:54:51 <fabiog> for Berlin either week 02/08 or 02/20
15:54:59 <rhochmuth> bmotz could you travel to berlin or san jose
15:55:00 <bklei> would love to go to berlin, will need to be really nice to my boss to get approval
15:55:29 <witek> Bogdan wanted to organize next one in Munich
15:55:34 <witek> in Sept
15:55:39 <fabiog> there is Cisco Live in Berlin the week of 02/15 so I will be traveling there anyway
15:55:45 <rhochmuth> yes, that was my preference too
15:55:46 <bmotz> berlin is probably easier to fit in the travel budget, but further afield isn't impossible
15:56:02 <rhochmuth> ok, thanks bmotz
15:56:20 <rhochmuth> we'llk start to need checking with our representative companies
15:56:27 <bklei> yup
15:56:45 <rhochmuth> can we get to the monasca-log-api in closing
15:56:47 <fabiog> I'll bring up the subject with Congress too
15:57:07 <tsv> rhochmuth: +1
15:57:10 <ddieterly> fabiog: can you talk to the president too? ;-)
15:57:26 <rhochmuth> #topic monasca-log-api
15:57:32 <rhochmuth> only 3 minutes
15:57:49 <tsv> 3 reviews :)
15:58:14 <fabiog> ddieterly: in progress ;-)
15:58:14 <witek> tomasz has some code ready, which partly covers same isues
15:58:20 <ddieterly> cool
15:58:47 <rhochmuth> didnt' realize you were working on it
15:58:51 <rhochmuth> too
15:59:25 <rhochmuth> how soon will your changes be post'd
15:59:25 <witek> i know, we still have internal reviews
15:59:43 <witek> have to consider to change the workflow
15:59:49 <witek> next days
15:59:59 <rhochmuth> how about move discussion to email then
16:00:03 <rhochmuth> we've run out of time
16:00:19 <rhochmuth> tsv witek and tomasz need to talk off-line
16:00:22 <fabiog> ciao!
16:00:24 <rhochmuth> or in irc
16:00:26 <witek> ok, ciao
16:00:30 <bklei> thx!
16:00:31 <tgraichen> bye
16:00:32 <ddieterly> ciao!
16:00:41 <rhochmuth> #endmeeting