15:00:54 #startmeeting monasca 15:00:55 Meeting started Wed Jun 27 15:00:54 2018 UTC and is due to finish in 60 minutes. The chair is witek. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:56 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:00:58 The meeting name has been set to 'monasca' 15:01:12 hello everyone 15:01:17 greetings 15:01:29 hi 15:01:39 hi joadavis and koji 15:02:09 Sweden and Germany watch football probably ;) 15:02:33 agenda: 15:02:38 https://etherpad.openstack.org/p/monasca-team-meeting-agenda 15:03:06 #topic reviews 15:03:33 Amir has submitted oslo implementation for log-api 15:04:07 that is the last step of the changes 15:04:14 https://review.openstack.org/#/c/578061/ 15:04:35 I have also updated our board: 15:04:43 https://storyboard.openstack.org/#!/board/60 15:05:35 we have still some changes with python 3 support 15:05:46 alembic 15:06:12 we should probably add Ceilosca to `Needs Help` 15:07:26 makes sense, unless we want to add a 'Blocked' column 15:08:06 before we move to deployment, two other short topics 15:08:20 #topic collaboration with Ceilometer 15:08:40 http://lists.openstack.org/pipermail/openstack-dev/2018-June/131443.html 15:09:01 I have offered help maintaining the Ceilometer repo 15:10:26 in return asked for merging our Monasca publisher and adding core-reviewer rights 15:10:38 they have very limited manpower 15:10:49 I'm curious if Telemetry team will accept the help or want to maintain control 15:11:08 I do worry that we have limits on our manpower too. :) 15:11:38 But it is definitely in OpenStack's interest to keep Telemetry going 15:11:38 that's true, and that's also the reason why we should work together 15:12:06 and leverage the work which has been done already 15:12:38 yes 15:13:10 I have started myself looking at changes in ceilometer repo 15:13:25 there is not much movement 15:13:57 it does get me thinking about using the Ceilometer service to gather events and send them to the Monasca Events API, rather than to Panko 15:15:18 yes, would we need a second publisher for this? 15:16:36 possibly, I haven't looked close enough to know if we could just route them through the existing Monasca Publisher code. 15:17:22 sounds good, please keep us updated 15:17:26 The ceilometer/events code would need to change - it does a fair bit of processing on events, which we would not need for Monasca Events (I think Ashwin pointed that out to me) 15:17:54 Probably worth some investigation and a spec 15:18:18 will you have time for that? 15:18:54 probably not this week, but I'll keep it on my list (Ashwin and I have another spec to write still) 15:19:35 we probably need another publisher, also we probably need to be able to post all the data in the incoming notification, right now it does some filtering to post only limited data (picks some metadata) 15:20:56 do we need all the data, or do you want to filter it in a different way? 15:22:14 anyway, it's great idea to have a spec for this 15:22:19 i think its best to post all the data, and do filtering on processing side 15:22:50 thanks joadavis and aagate, please keep us updated 15:23:12 #topic ptg 15:23:20 http://lists.openstack.org/pipermail/openstack-dev/2018-June/131689.html 15:23:46 the initial list of sessions for the next PTG has been published 15:24:03 or rather projects and sigs 15:24:50 for those having budget problems I encourage to apply for Travel Support 15:25:35 Do we need to feed back if there are any projects or sigs we don't want to overlap with so we can work together? 15:25:44 witek: as operator can i attend this session ? 15:26:07 pandiyan_: I think you are welcome any time. :) 15:26:07 joadavis: yes, but there is no agenda yet 15:26:25 pandiyan_: yes, sure it's also for operators 15:26:38 because our setup is really big and it will be great to attend this session, facetime always benefits mutually. 15:26:50 thinking of our interactions last PTG, I don't see Congress or Vitrage on the list. 15:27:21 Erik from Congress will attend, but they won't have own sessions 15:29:11 #topic deployment 15:30:14 pandiyan_: it's yours 15:30:34 Hi All, In our organisation, we have tried to change monasca deployment to scale it, but kafka cluster doesn't support monasca application, as level1 we are doing it out on singel node setup 15:31:36 coming to topic, we are trying to deploy monasca without keystone image and use openstack keystone details in respective image environment variables as like http://paste.openstack.org/show/724392/ 15:32:09 as per that, monasca api is running fine, where grafana is failing with keystone authentication 15:32:51 Here is my keystone entry for addning monasca service http://paste.openstack.org/show/724410/ where "monasca-user" role is added 15:33:36 let me know whether any changes need to be done in my docker compose file 15:33:48 am done with my questions 15:34:36 witek: looking forward people here to look and suggest me 15:35:30 firstly, it's in general not true that Monasca does not support clustered Kafka 15:35:46 monasca-docker does not support it 15:36:11 but it's a matter of configuration on components 15:37:02 witek: hope you agree for production enviroment good to go with scalablity 15:37:42 e.g. take a look at this template: https://gitlab.com/monasca-installer/ansible-monasca-api/blob/master/templates/api-config.yml.j2#L22 15:38:21 we have configured kafka zookeeper on cluster setup on baremetals, where monasca failed to detect kafka master, sometimes monasca-notification or other services throwing error as "zookeeper not able to find leader" 15:38:37 with docker setup without cluster working fine 15:39:27 you have to provide own configuration if you want to run in cluster, it's not included in monasca-docker 15:40:00 okay, sure will do. even i looking into high level setup which you have shared me last meeting 15:40:16 coming to this grafana issue with keystone, may i know how to ? 15:41:14 so, you have user 'monasca' in 'monasca' project 15:41:23 yes 15:41:38 as per keystone entry 15:41:50 the user has the role 'monasca-user' which is authorized to query the API 15:42:30 yes, i have created monasca-user role and added to monasca user which is added to monasca project 15:43:00 http://paste.openstack.org/show/724410/ 15:43:06 so you can configure this user for grafana-init container 15:43:08 should work 15:43:44 i didn't get that, can i have enviroment variables 15:44:16 GRAFANA_USERNAME: monasca 15:44:28 no, wrong one, sorry 15:46:02 pandiyan_: can you provide the error message of authentication failure in grafana? 15:47:31 http://paste.openstack.org/show/724414/ 15:48:20 and please share `grafana-init` container logs 15:48:22 koji, above error 15:48:34 if possible. 15:48:40 that should be GRAFANA_USERNAME: monasca; GRAFANA_PASSWORD: password 15:49:51 http://paste.openstack.org/show/724415/ 15:50:24 haruki: above details 15:51:09 thank you 15:52:18 grafana-init cannot reach http://grafana:3000/login 15:52:56 witek: didnt get you 15:53:34 grafana-init is there to install Monasca datasource and dashboards 15:53:53 yes, as witek assumed, grafana-init fails to create datasource and default dashboard because of keystone authentication error. 15:53:53 is this part of the grafana log edited ? `userId=0 orgId=0` 15:54:30 how to resolve it, can someone help,, i have shared info 15:55:00 as witek suggested created GRAFANA_USERNAME: monasca; GRAFANA_PASSWORD: password 15:57:25 can you open Grafana login page? 15:58:29 https://github.com/monasca/monasca-docker/tree/master/grafana-init#configuration 15:59:25 witek: http://paste.openstack.org/show/724416/ 16:00:16 http://paste.openstack.org/show/724417/ 16:00:17 you must configure port forward for grafana image 16:00:33 c7d098dd3b68 monasca/grafana:4.0.0-1.4.2 "/start.sh" 4 hours ago Up 4 hours (healthy) 0.0.0.0:3000->3000/tcp monascadocker_grafana_1 16:00:36 port forward is 3000 in docker compose file 16:00:43 maybe, you need to specify admin endpoint for GF_AUTH_KEYSTONE_AUTH_URL, but i cannot confirm it now 16:01:02 have to close the meeting now 16:01:07 Currently, I have not set up monaca-docker original repo, but it seems some configuratiosn need to integrate with external keystone, `maybe` 16:01:09 see you next week 16:01:10 bye 16:01:14 koji: as per my compose file its pointing to to that only 16:01:17 okay 16:01:19 #closemeeting 16:01:30 bye 16:01:38 #endmeeting