15:01:18 <witek_> #startmeeting monasca
15:01:19 <openstack> Meeting started Wed Nov 20 15:01:18 2019 UTC and is due to finish in 60 minutes.  The chair is witek_. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:01:20 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:01:23 <openstack> The meeting name has been set to 'monasca'
15:01:26 <witek_> hello everyone
15:01:33 <chaconpiza> hi
15:01:44 <witek_> hi chaconpiza
15:02:26 <witek_> is it just two of us today?
15:02:46 <witek_> the agenda:
15:02:50 <witek_> https://etherpad.openstack.org/p/monasca-team-meeting-agenda
15:03:10 <witek_> #topic new meeting time
15:03:13 <chaconpiza> today only me from fj
15:03:27 <witek_> the results of the poll are here
15:03:29 <witek_> https://doodle.com/poll/ey6brvmbsubkxpp9
15:03:41 <dougsz> hey there
15:03:46 <witek_> hi dougsz
15:03:50 <chaconpiza> hi
15:04:48 <witek_> according to the poll the best new time for the meeting would be on Tuesdays, 1pm UTC
15:05:04 <witek_> is that OK for you?
15:05:39 <dougsz> Works for me
15:05:56 <chaconpiza> Works for me too
15:06:05 <witek_> let's move it then
15:06:13 <witek_> thanks for voting
15:06:57 <witek_> I'll update the overview page and send an email
15:07:19 <witek_> #topic dropping Python 2 support
15:07:52 <witek_> TC has decided on dropping Python 2 support in Ussuri cycle
15:08:09 <witek_> there are already changes removing unit tests
15:08:17 <witek_> https://review.opendev.org/#/q/(projects:openstack/monasca+OR+project:openstack/python-monascaclient)+is:open+topic:drop-py27-support
15:08:57 <witek_> it has also been nominated to the Community Wide Goal
15:09:01 <witek_> https://governance.openstack.org/tc/goals/selected/ussuri/drop-py27.html
15:09:53 <witek_> taking a look at the changes I noticed that our tox configurations are not always correct
15:10:19 <witek_> and so for examples py37 tests are not run in some cases
15:10:49 <witek_> we should take this occasion and review our tox.ini files
15:11:04 <chaconpiza> As well we continue using PY2 to build Grafana in devstack:
15:11:05 <chaconpiza> https://github.com/openstack/monasca-api/blob/master/devstack/plugin.sh#L1266
15:12:01 <witek_> good point, has anyone checked recently if it can be done with Python 3?
15:12:26 <chaconpiza> no
15:12:46 <dougsz> I wonder if it is worth the effort
15:13:16 <chaconpiza> we had two init scripts that were only PY2 compatible but recently adriancz fixed them.
15:13:36 <witek_> that's still the old forked Grafana, right?
15:14:19 <chaconpiza> https://github.com/openstack/monasca-api/blob/dfd942352611cb4aba53f6bb3d9628ac70b085c7/devstack/settings#L159
15:14:54 <dougsz> `GRAFANA_REPO=${GRAFANA_REPO:-"https://github.com/monasca/grafana.git"}` yeah
15:16:43 <witek_> ideas how to move forward?
15:17:20 <dougsz> It seems like putting effort into app credentials + getting rid of the fork would be a good objective
15:17:46 <witek_> +1
15:18:17 <joadavis> +1
15:18:18 <chaconpiza> what for this fork was created?
15:18:28 <witek_> Keystone authentication
15:18:29 <joadavis> to include keystone integration
15:18:34 <chaconpiza> ah ha
15:19:17 <joadavis> heh, sorry. I'm playing catch up as my laptop froze 10 minutes before the meeting and I switched laptops
15:19:43 <witek_> I think having python 2 here now will not break gate jobs
15:19:59 <witek_> so we can work on app credentials in this cycle
15:20:04 <witek_> and upgrade when ready
15:20:07 <joadavis> right, I don't think any gates test grafana integration
15:20:13 <dougsz> sounds good
15:21:33 <witek_> Cassandra doesn't have Python 3 jobs
15:22:07 <witek_> meaning it won't be tested anymore after migration on master
15:22:18 <witek_> only on stable branches
15:22:44 <joadavis> yuck.  that is worth fixing if we can
15:23:33 <witek_> anything else we have to consider?
15:24:24 <witek_> #topic logs gate job
15:25:26 <witek_> http://zuul.openstack.org/builds?job_name=monascalog-python3-tempesthttp://zuul.openstack.org/builds?job_name=monascalog-python3-tempest
15:25:37 <witek_> the failure rate is scary
15:26:08 <witek_> this job blocks many reviews right now
15:26:15 <dougsz> yeah, I had a quick look but didn't get to the bottom of it yet
15:27:09 <witek_> I'll try to find out if we can run it on a bigger vm
15:28:20 <witek_> do we want to create an etherpad to collect debug information?
15:29:12 <dougsz> Good idea
15:29:30 <dougsz> I should have some time early next week to help debug
15:29:48 <witek_> https://etherpad.openstack.org/p/monascalog-python3-tempest
15:30:28 <witek_> let's put there all we know
15:31:34 <witek_> #topic reviews
15:31:53 <witek_> we've managed to merge few changes this week
15:32:00 <witek_> I've updated the board
15:32:14 <witek_> https://storyboard.openstack.org/#!/board/190
15:32:41 <witek_> still few changes needing reviews
15:32:50 <witek_> https://review.opendev.org/#/q/topic:story_2006059
15:33:28 <witek_> this is for buffering messages in API
15:33:28 <dougsz> Thanks for those patches witek
15:34:05 <dougsz> As you say, all blocked on tempest failures
15:34:22 <witek_> I've added possibility to pass options to the new Kafka producer
15:34:54 <witek_> it can control number of messages in buffer and many other values
15:35:17 <witek_> https://review.opendev.org/674204
15:35:35 <witek_> this one fixes gate jobs for events
15:35:51 <witek_> https://storyboard.openstack.org/#!/story/2006783
15:35:57 <witek_> https://storyboard.openstack.org/#!/story/2006783
15:36:06 <witek_> https://storyboard.openstack.org/#!/story/2006837
15:36:17 <witek_> these are for periodic notifications
15:36:30 <witek_> I haven't had time to look at these yet
15:37:12 <witek_> and then IPv6 change from gmann
15:37:17 <witek_> https://review.opendev.org/673274
15:37:55 <witek_> I guess he's busy with Python 2 dropping now, we should help on this one
15:39:08 <sum12> https://review.opendev.org/#/c/677943/ this too
15:40:15 <sum12> is also for IPV6
15:40:17 <joadavis> did yeey (Guang) contact you about backporting his https://review.opendev.org/#/c/690477/ back as far as possible?
15:40:32 <witek_> than you sum12
15:40:50 <witek_> joadavis: no
15:40:52 <joadavis> he would like to take it to pike, but I talked to him yesterday about how that is difficult with EM releases
15:41:47 <witek_> I think it will be easier to create a patch in packaging
15:42:13 <joadavis> I think that is where my chat with yeey ended up
15:42:28 <joadavis> but it doesnt hurt to ask. :)
15:42:36 <witek_> no, it doesn't :)
15:43:30 <witek_> I saw a new patch on ELK upgrade change
15:43:49 <witek_> https://review.opendev.org/#/c/675061/
15:44:04 <witek_> chaconpiza: is it functional now?
15:44:20 <chaconpiza> not sure, I will ask arseni tomorrow
15:45:02 <witek_> thanks, he can leave a comment on review if it's ready
15:45:20 <witek_> pinging in channel also works
15:46:14 <witek_> any other reviews?
15:46:17 <sum12> https://review.opendev.org/#/c/679225 I myself forgot about this one until someone told me that it is needed I will remove the WIP but please keep on eye on it
15:47:07 <witek_> thanks, adding to the list
15:47:52 <sum12> thanks
15:48:07 <witek_> #topic AOB
15:48:21 <chaconpiza> what was the motivation to change the index-version in influxdb?
15:48:21 <chaconpiza> Now we have in devstack Influx 1.7.6 index-version = "tsi1"
15:48:21 <chaconpiza> In the past was used index-version = "inmem"
15:49:00 <dougsz> chaconpiza: tsi1 is recommended now
15:49:10 <chaconpiza> does it fit better with metric pipeline?
15:49:36 <chaconpiza> Our influxdb 1.3.4 is consuming too much memory almost 44% of 32GB RAM with no reason, logs look ok. Just normal stuff and every day a shard is being deleted as expected.
15:49:39 <witek_> it allows for storing more metrics
15:49:43 <dougsz> it reduced memory requirements - it used a disk based index which is cached into RAM
15:50:20 <chaconpiza> I found as well in influxdb release notes some fixes about memory leaks in 1.4.x versions
15:50:34 <dougsz> I would skip 1.4 entirely and go straight to 1.7
15:50:45 <dougsz> There is a migration procedure in the docs
15:51:11 <dougsz> We've had no issues with the migration
15:51:11 <chaconpiza> are you using 1.7 in prod?
15:51:14 <dougsz> yeah
15:51:18 <chaconpiza> cool
15:51:46 <dougsz> it's default in Kolla Ansible
15:54:08 <witek_> from other topics, there is an interesting discussion on openstack-discuss
15:54:13 <chaconpiza> thanks, we will test the migration to tsi1
15:54:15 <witek_> http://lists.openstack.org/pipermail/openstack-discuss/2019-November/010922.html
15:55:52 <witek_> if there are no other topics, I'll be wrapping up
15:56:22 <dougsz> ok, thanks witek, thanks all
15:56:29 <witek_> thanks for good discussions
15:56:31 <chaconpiza> thanks all
15:56:39 <sum12> thanks
15:56:40 <witek_> next week we meet on Tuesday, at 1pm
15:56:44 <witek_> UTC
15:56:48 <chaconpiza> yeah
15:57:11 <witek_> bye, bye
15:57:16 <witek_> #endmeeting