15:00:21 <witek> #startmeeting monasca
15:00:22 <openstack> Meeting started Wed Jan  2 15:00:21 2019 UTC and is due to finish in 60 minutes.  The chair is witek. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:23 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:00:25 <openstack> The meeting name has been set to 'monasca'
15:00:37 <witek> Hello
15:01:11 <witek> hi joadavis
15:01:43 <joadavis> good morning/afternoon
15:02:05 <witek> welcome in 2019
15:02:14 <joadavis> yes, happy new year. :)
15:02:24 <witek> is anyone else around?
15:02:50 <witek> let's start
15:03:00 <witek> #topic Vitrage integration
15:03:08 <witek> just a short update
15:03:43 <witek> Ifat has submited a spec for Monasca datasource
15:03:51 <witek> https://review.openstack.org/627180
15:04:09 <witek> I didn't have time to read it yet, though
15:04:36 <joadavis> I'll take a look today
15:04:57 <witek> I'll do in next days as well
15:05:48 <witek> apart from that I don't have any other topics
15:06:07 <witek> joadavis: anything from you?
15:06:22 <joadavis> I don't have anything new.
15:06:41 <joadavis> Before my holiday break, I was chasing an old bug again
15:07:26 <joadavis> in monasca-agent, it occasionally can produce a metric with None value, which causes an error in mon-api and gets requeued
15:07:55 <joadavis> I've attempted a few fixes in the past to try to narrow it down. I'll be retesting my last attempt this week
15:08:35 <joadavis> I'm going to try to cram a lot in to this week. :)
15:08:41 <witek> actually the error status should return enough information to drop the measurement
15:09:08 <witek> and not queue it again
15:09:23 <joadavis> yes, unfortunately the http error code is lost between the response processing and the validation code. So it doesn't know the difference between a 500 and a 422
15:09:43 <joadavis> I could parse that back out of the exception, but the separate code is not passed back up
15:09:55 <joadavis> so it just requeues the whole list of metrics
15:11:13 <joadavis> The odd thing is, from my previous investigation the code that produces the metric (an http check) should have the value hard coded to 0 or 1 and never produce a None
15:12:39 <witek> I see, perhaps we should change the logic and queue the failed messages individually and not in batch?
15:13:08 <joadavis> possibly.
15:13:16 <joadavis> the story is https://storyboard.openstack.org/#!/story/2004264
15:13:46 <witek> thanks
15:14:12 <witek> OK, if we don't have anything else, I'll close now
15:14:52 <witek> tomorrow we have another meeting at 7 UTC
15:14:57 <witek> thanks joadavis
15:15:02 <witek> see you next week
15:15:05 <joadavis> thank you
15:15:24 <witek> #endmeeting