15:00:17 #startmeeting monasca 15:00:18 Meeting started Wed Jul 10 15:00:17 2019 UTC and is due to finish in 60 minutes. The chair is witek. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:19 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:00:21 The meeting name has been set to 'monasca' 15:01:00 hello everyone 15:01:09 Hello 15:01:10 o/ 15:01:12 hello 15:01:20 hi 15:01:24 hi 15:01:28 hi 15:01:31 few topics in the agenda today 15:01:34 https://etherpad.openstack.org/p/monasca-team-meeting-agenda 15:01:41 please add if you have any 15:01:59 # topic Faust library evaluation 15:02:03 #topic Faust library evaluation 15:02:31 I've spent some time this week playing with the Faust library 15:02:44 https://faust.readthedocs.io 15:03:10 I've collected most important finding in this etherpad: 15:03:12 https://etherpad.openstack.org/p/faust-evaluation 15:03:33 there are few issues 15:03:49 like libraries not common in OpenStack 15:04:03 or forked version of aiokafka library 15:04:40 from the technical point of view I haven't found any blocker though 15:05:08 one issue might be, that they have rather small community 15:05:29 most commits are from the single person 15:05:53 but the community chat seems to be quite lively 15:06:11 so this is as an alternative to using prometheus? 15:06:54 this is a library which could be used to write new thresholding engine 15:07:00 a, ok 15:07:13 https://github.com/witekest/faust-examples 15:07:47 here I have written a simple example of how the aggregation functions on hopping window could be implemented 15:08:17 similar to what we have in monasca-thresh right now 15:09:02 theoretically, the library could be used to write the aggregation engine as well 15:10:25 Interesting, thanks for collecting your ideas witek 15:11:52 I know Matthias was having a look at this library already as well 15:13:12 OK, I guess that's all for this topic, if there are no questions 15:13:56 #topic alarm expression parser 15:14:26 chaconpiza: your stage 15:14:34 Hi, I noticed that we have a duplication of code 15:14:40 https://opendev.org/openstack/monasca-api/src/branch/master/monasca_api/expression_parser/alarm_expr_parser.py 15:14:42 and 15:14:48 https://opendev.org/openstack/monasca-common/src/branch/master/monasca_common/expression_parser/alarm_expr_parser.py 15:15:02 The code of alarm_expr_parser was copied from monasca-api to monasca-common in order to be used in monasca-ui 15:15:16 But after more than a year it isn't used in monasca-ui 15:15:43 Here the leading story: https://storyboard.openstack.org/#!/story/2001837 15:16:27 Unfortunately the owner, Lukasz Zajaczkowski is not active anymore in Monasca 15:16:50 I would leave it in monasca-api and remove from monasca-common 15:18:32 any opinions? 15:18:59 I 15:19:01 if we don't want to use it in ui then fine 15:19:43 +1 15:19:44 it makes sense to use it in the UI but apparently wasn't too important 15:19:59 Yeah, i guess it's only for usability 15:20:31 Sounds reasonable to remove it from common 15:20:47 I'm fine with proposed approach, if someone wants to work again on it, we can move it again 15:22:02 I've seen Adrian and chaconpiza refactoring changes 15:22:12 chaconpiza: will you take care of this one as well? 15:22:16 sure 15:22:22 thanks 15:22:48 #topic Zuul flakiness 15:23:00 that's my 15:23:18 I'm really starting to hate zuul 15:23:26 :D 15:23:34 why? :) 15:23:49 if I understood correctly zuul have too weak machines for testing monasca 15:24:19 and that's why we see my timeouts in tempest tests 15:24:41 I think it's just a standard vm with 8 GB RAM 15:24:47 I'm starting to hate writing recheck 15:25:23 it's too small for all services we are using 15:26:27 does openstack run any tests with bigger machines/vm? 15:26:34 don't know 15:27:20 hmm, I can check on what Kolla use and report back 15:27:32 there are some pretty big deploys in the CI there 15:27:35 thanks dougsz, that would help 15:27:58 because now irrelevant changes could be blocked for few days from merging because of timeouts... 15:28:03 we could also get rid of some jobs which are constantly failing 15:28:34 do you consider that a fix? 15:28:58 no, but it's wasting resources 15:29:38 and a lot of time 15:30:56 nothing more from me about this 15:30:57 another option would be to check, if we can trim out our setup in DevStack plugin in terms of memory consumption 15:31:58 I think this is quite difficult 15:33:44 it has been reported recently in openstack-discuss that other projects also have similar issues 15:35:35 OK, I'll wait for input from dougsz and check with infra team if we can use larger flavors 15:36:31 other ideas? 15:37:11 no idea, we are using few heavy services 15:38:18 there are some logs available in CI which could help debugging 15:38:24 like: dstat-csv_log.txt.gz 15:38:49 or dstat-csv_log.txt.gz 15:39:13 or screen-peakmem_tracker.txt.gz 15:40:01 #topic mid-cycle 15:40:26 thanks for voting in the doodle list 15:40:36 https://doodle.com/poll/zszfxakcbfm6sdha 15:40:48 no date is perfect for everyone 15:42:00 we could choose Mon 22, or Wed 24 15:43:00 I hope joadavis can do it on Wed, otherwise I'd go for Monday 15:43:40 I'll let you know on the mailing list and IRC 15:43:55 and send the invitation 15:44:27 please also start filling in the agenda 15:44:29 https://etherpad.openstack.org/p/monasca-train-midcycle 15:44:47 the format is not strict, please feel free to add any topics 15:45:04 that's all from me 15:45:46 any other topics for today? 15:46:00 sumitjami proposed openstack/monasca-agent master: WIP Enable unit tests for py36 https://review.opendev.org/657829 15:46:42 if not, I'll be wrapping up 15:46:52 thanks for joining 15:46:58 thank you 15:47:01 and see you next week 15:47:25 #endmeeting