15:00:38 <witek> #startmeeting monasca
15:00:40 <openstack> Meeting started Wed Apr 25 15:00:38 2018 UTC and is due to finish in 60 minutes.  The chair is witek. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:41 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:00:44 <openstack> The meeting name has been set to 'monasca'
15:00:59 <witek> Hello everyone!
15:01:19 <jgrassler> Hello
15:01:28 <koji> hello
15:01:30 <haruki> hi
15:01:46 <aagate> hello!
15:01:54 <joadavis> greetings!
15:02:08 <witek> agenda for today:
15:02:12 <witek> https://etherpad.openstack.org/p/monasca-team-meeting-agenda
15:02:32 <witek> #topic next ptg
15:03:02 <witek> I have been asked if Monasca will participate in the next PTG
15:03:15 <witek> it will take place September 10-14, 2018, Denver, Colorado
15:04:17 <witek> the deadline for reply is quite short again, so I have sent the email to mailing list
15:04:49 <witek> with the link to Google form
15:05:03 <witek> asking what is your opinion on joining the PTG
15:05:10 <witek> https://goo.gl/forms/nWs4LX0MMBKZOCu72
15:05:29 <tobiajo> hi
15:05:51 <witek> hi, I see there are apparently some network problems
15:06:55 <witek> I've got already some responses from SUSE
15:07:12 <witek> what do other folks think about joining the PTG?
15:07:41 <dougsz> filled in the form - I like the idea of joining in principle, but it's a long way and will be expensive for us.
15:08:24 <witek> thanks dougsz, yes travel is definitely a problem
15:09:21 <tobiajo> we at op5 haven't got any direction yet. in my personal opinion i feel that it's a too far away if we all are situated in europe
15:10:14 <jgu_> it is just a couple of us here in U.S.
15:12:43 <witek> another argument was, if we'll have enough topics to cover
15:13:08 <witek> timothyb89: would you be attending?
15:14:52 <witek> OK, please leave your comments in the form, we'll have to decide soon
15:15:18 <witek> having remote meeting is also an alternative
15:15:48 <witek> #topic log query API
15:16:28 <witek> doug, I have added your reviews
15:16:36 <dougsz> So we have a prototype I mentioned at the last PTG, and after looking at the code, there is a lot of common code shared with the Monasca-API
15:16:40 <dougsz> thanks witek
15:17:01 <dougsz> The first two changes are to factor out common methods to make the log-query-api patch as small as possible.
15:17:37 <tobiajo> do we want much shared code in common? some would call it a "distributed monolith" due to the coupling
15:18:29 <dougsz> I think some things should work consistently across both APIs, like role based authentication
15:20:20 <witek> I agree with dougsz here, and that's not that much, what get's extracted to monasca-common
15:20:48 <tobiajo> for this patch i don't have any objection
15:21:25 <dougsz> thanks for checking tobiajo - i see your point in general
15:21:26 <tobiajo> but i still think there is too much shared in general
15:22:21 <dougsz> Which parts would you factor out from common into the APIs?
15:23:15 <tobiajo> sorry i was off topic, meant in general, i have nothing to add in this specific topic
15:23:34 <dougsz> np, thanks
15:24:39 <dougsz> But I guess for these patches - if people are generally happy with extracting that code from common I can polish them up.
15:24:53 <dougsz> And get to work on submitting the actual log-query-api change
15:25:28 <dougsz> sorry s/from common/to common
15:25:53 <witek> apart from stuff for APIs, we also have Kafka code there
15:26:33 <dougsz> in the common lib?
15:26:45 <tobiajo> yes monasca-common
15:27:39 <tobiajo> a shared wrapper around the kafka client, and the forked kafka client itself...
15:27:39 <dougsz> I noticed some inconsistencies across the APIs with Kafka, I guess it would be nice if they were configured in a standard way, but that's for another change.
15:28:46 <witek> tobiajo: do you think we should take Kafka wrapper to individual components?
15:30:13 <dougsz> I almost wonder the opposite in that maybe the Monasca APIs should be merged into one - but again, something for another time!
15:30:36 <tobiajo> witek: yes. and is a kafka wrapper even necessary?
15:33:04 <witek> we'll have to think of it when upgrading Kafka client
15:33:13 <tobiajo> i think a kafka wrapper is a good idea in theory, but how easy is it to keep it up to date?
15:34:17 <dougsz> I wonder if anything can be used from Oslo - I know they are using it for OpenStack notifications
15:36:22 <witek> I think they use python-kafka
15:36:32 <witek> and we want to move to pykafka or confluent
15:37:01 <dougsz> thanks, i'll have to take a closer look when I get a chance.
15:37:18 <tobiajo> dougsz: this impl? https://review.openstack.org/#/c/189006/
15:38:40 <dougsz> tobiajo: Haven't seen that - I just know they were asking about it at the PTG and spent a bit of time reviewing the Kafka patches for Kolla
15:39:11 <witek> anyway, let's take care about what we're extracting to common
15:39:28 <witek> and folks, have a look at dougsz changes
15:40:01 <witek> I think we can move on, right?
15:40:14 <dougsz> thanks witek, sounds good.
15:40:17 <tobiajo> y
15:40:19 <witek> thanks
15:40:43 <witek> #topic token verification
15:40:56 <witek> tobiajo: that's yours?
15:41:02 <tobiajo> y
15:41:12 <tobiajo> for reference: https://storyboard.openstack.org/#!/story/2001913 & https://storyboard.openstack.org/#!/story/2001914
15:41:41 <tobiajo> i added this now a little late, and it could be the case that we are not doing it right
15:41:54 <witek> I guess we're just using Keystone middleware for that
15:42:33 <dougsz> same for us
15:43:23 <tobiajo> i have ran into an issuess where monasca-api gets more open connections to memcached than should be needed. and we suspect that these connections is not handled properly
15:44:18 <witek> we have custom filter in monasca-api
15:44:19 <tobiajo> have someone else had this issue? if not, should offline token verification work?
15:44:25 <witek> https://github.com/openstack/monasca-api/blob/master/monasca_api/healthcheck/keystone_protocol.py
15:45:18 <witek> it's used not to authorize requests to healthcheck endpoint
15:45:56 <witek> are you using it?
15:46:20 <witek> https://github.com/openstack/monasca-api/blob/master/etc/api-config.ini#L11
15:47:27 <tobiajo> how would that be related to offline verification of fernet tokens?
15:50:44 <witek> it should work just as described here: https://docs.openstack.org/keystonemiddleware/latest/middlewarearchitecture.html
15:50:45 <tobiajo> i don't want to skip auth, the Keystone middleware should be able to support offline verification. maybe we are just doing it wrong?
15:54:01 <witek> I'm afraid I cannot help right now, could you add relevant information about your configuration to the story?
15:54:49 <tobiajo> too bad, but i guess you others have offline token validation?
15:55:13 <witek> memcached
15:55:23 <tobiajo> i see
15:55:30 <tobiajo> that doesn't scale
15:55:45 <dougsz> tobiajo - i'd assumed so, but I haven't verified it. Our config is here: https://review.openstack.org/#/c/559375/
15:56:00 <tobiajo> dougsz & witek: thanks!
15:56:02 <dougsz> (memcached as well)
15:56:40 <witek> #topic ceilometer publisher
15:57:03 <witek> joadavis: I've seen your change, thanks
15:57:09 <joadavis> Ah. Just wanted to give a quick update. :)
15:57:28 <joadavis> I think all needed changes are in, including some feedback from Gordon Chung
15:57:38 <joadavis> Now we just wait for reviews and approval
15:57:54 <witek> I've seen the comment, let's hope we can get it in
15:58:59 <witek> OK, thank you everyone for joining and for the discussion
15:59:13 <witek> see you next week
15:59:18 <witek> bye bye
15:59:26 <dougsz> Thanks!
15:59:26 <koji> thx
15:59:30 <haruki> bye
15:59:31 <jgrassler> See you!
15:59:31 <koji> bye
15:59:31 <tobiajo> bye
15:59:38 <aagate> thx! bye
15:59:46 <witek> #endmeeting