15:02:12 #startmeeting horizon 15:02:13 Meeting started Wed Jan 9 15:02:12 2019 UTC and is due to finish in 60 minutes. The chair is e0ne. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:02:14 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:02:16 The meeting name has been set to 'horizon' 15:02:20 o/ 15:02:34 rdopiera: hi 15:02:39 #info 15:02:44 0/ 15:02:47 amotoki: hi. are you around? 15:02:57 hi 15:03:21 great. looks like we're OK to start today 15:03:43 I hope, you enjoyed Christmas and New Year holidays 15:03:52 hello all 15:04:02 #link https://wiki.openstack.org/wiki/Meetings/Horizon#Agenda_for_Next_Meeting 15:04:40 #topic Notices 15:05:12 OpenStack Summit CFP is open until Wednesday, January 23 at 11:59pm PT. 15:05:37 don't forget to submit a talk proposal if you're going to do it 15:06:36 #link https://www.openstack.org/summit/denver-2019/call-for-presentations/ 15:07:24 OpenStack PTG will be in the same location right after the summit 15:07:27 #link https://www.openstack.org/ptg 15:07:55 amotoki, rdopiera: are we OK to have one day for the PTG? 15:08:19 e0ne: I can't come anywyas, so it's fine by me 15:08:28 rdopiera: :( 15:09:18 I have to answer until January 20th if we want to get some room for PTG discussions 15:10:58 I hope, we'll have some good discussions there 15:11:11 amotoki: are you going to attend the Summit and PTG? 15:13:12 I haven't decided and got approved yet, but i hope so 15:13:49 I didn't get an approval too, but I have some time for it 15:14:39 I'll sent a note to organizers that we're going to attend it. I think we can cancel if if nobody goes 15:15:05 sounds good 15:15:20 #topic New ceilometer-dashboard repository 15:15:42 jakub: could you please share with us what are you proposing? 15:17:22 #link https://review.openstack.org/#/c/619235/ 15:17:29 #link https://bitbucket.org/ultimumtechnologies/ceilometer-dashboard/src 15:18:57 so Telemetry team are going to create own plugin for Horizon 15:19:00 it's great 15:19:03 It seems we need a discussion on naming per Doug's comment 15:19:10 We would like to add a plugin to horizon for ceilometer. Ceilometer stores data in gnocchi and these data can be displayed in Grafana. But if you want to see it you have to login to Grafana and create some basic structure. And this is the point, our hoziron plugin creates this structure via a button and also can redirect you to the grafana dashboar without need of additional authentication. 15:20:20 We would like to get the code under Openstack. 15:21:00 Regarding naming - as I just described above, it is not a grafana plugin, therefore there is no need to rename the repository 15:21:19 A dashboard for telemetry is a missing point for a long time :) 15:21:53 jakub: I see. AFAIU, it is a horizon plugin which leverages/redirects to grafana UI 15:22:09 is it right? 15:22:24 basically yes 15:22:48 but not only 15:23:26 ack 15:23:44 from the governance perspective, we need an ack from the telemetry PTL 15:23:52 amotoki: +1 15:24:35 we use plugin model now. this means dashboard supports for non-core features are hosted by corresponding projects. 15:24:58 do we see any blocking issue from horizon perspective? 15:25:10 nothing I'm familiar with 15:25:57 I haven't followed comments/discussions in the proposed review yet. I will look thru it tomorrow. 15:25:58 jakub asked me if horizon team wants to share responsibility to maintain this plugin 15:26:23 in general, we don't have capacity to maintain plugins too 15:27:23 but I think it's reasonable if horizon team will have +2 on plugins for some emergency cases .e.g. if gate is blocked by some horizon or dependency changes 15:27:41 IMO, that's all we can do 15:27:42 does "share" mean that both the telemetry and horizon teams co-maintain it? 15:28:23 e0ne: yes I have asked you because PTL of Telemetry did not want to share the responsibility 15:28:57 jakub: it's up to your team 15:28:59 jakub: does it mean that the telemetry team cannot host it? 15:29:35 just a note, I am not member of Telemetry team 15:29:40 jakub: or would the telemetry team like to maintain it by themselves? 15:30:14 "share" means some team shares some reponsibility with another team 15:30:22 so I am a bit confused with the statement. 15:30:43 amotoki: yes, AFAIK they do not want to host it 15:30:49 hm... 15:30:55 jakub: so who will maintain it? 15:31:23 our company, if is that what you are asking 15:32:45 I'm confused now 15:32:47 this sounds a governance discussion. we need to consider both maintenance bandwidth and which project can host it. 15:32:58 amotoki: +1 15:33:31 re: the maintenance bandwidth, jakub's company can provide it (at least for now) 15:33:36 the question is also whether it needs to be official or unofficial 15:33:48 if there is any difference 15:34:08 re: the project governance there are several choces: horizon official , telemetry official and unofficial 15:35:04 re: the diffrence between offical vs unofficial, at least contributors to official projects get ATC status 15:35:45 in addition, "official" would be a good reason to contribute such project more compared to "unofficial" ones. 15:36:42 it would be good it telemetry project can host it 15:37:03 I think it is better to raise this to TC with what we discussed so far. 15:37:08 I'm not sure that we should host it under Horizon umbrella 15:37:46 including the fact that the telemetry PTL would like not to host it (in personal conversation?) 15:38:04 re: yes, via email 15:38:05 AFAIK, this is the first case that any project would like not to host its dashboard. 15:38:15 amotoki: :( 15:38:34 we do not insist on the official.. originaly I went unofficial but as you can see in the change request Andreas Jaeger wanted me to create it under governance so that is why I am here now 15:39:07 we simply wanted to be somehow under openstack not necessarily official 15:39:36 jakub: yeah, generally speaking it is a good suggestion, but the situation seems a bit complicated..... we didn't expect it :( 15:39:50 amotoki: me neither :) 15:39:58 so what is the next step? 15:40:19 that is what I would like to know as well 15:40:41 I see two ways: (1) raise this situation to TC, or (2) create a repo as unofficial project. 15:41:27 how can I do the (2) ? 15:42:11 jakub: it is simple. what we need to do is to send a mail explaining the current situation to openstack-discuss ML with [tc] tag. 15:42:30 jakub: here is some manual https://docs.openstack.org/infra/system-config/unofficial_project_hosting.html 15:43:02 you can also join TC meetings to discuss it but I believe sending a mail to the ML is a good first step 15:43:03 amotoki: what is ML? 15:43:12 I see now 15:43:12 jakub: openstack-discuss ML 15:43:14 jakub: mailing list 15:43:28 ok then, I will do that 15:43:52 aha, ML was a common abbrev of mailing list, but it now means machine learning :) 15:45:03 jakub: when you send a mail, could you iinclude "[tc][telemetry][horizon]" in the subject? 15:45:10 this is what I mean by "tag". 15:45:20 sure, I can 15:45:53 we can support the discussion then. 15:46:12 ok, thank you 15:46:34 jakub: thanks for raising this topic 15:47:24 can we move forward? I've got one more topic to discuss 15:47:26 thanks 15:48:06 #topic unused XStatic-* packages 15:48:44 we've got few xstatic-* libraries which we don't use in Horizon and they were not released at all 15:48:48 e.g. https://github.com/openstack/xstatic-angular-ui-router 15:49:31 do we want to drop them from horizon deliverables? 15:50:05 such repos were created about two years ago and we still don't use them 15:50:15 I am okay to drop them unless any horizon plugins depend on it 15:50:39 s/it/them/ 15:51:00 amotoki: I don't think somebody uses them because we don't have any releases 15:51:10 but I'll double check before removing them 15:51:13 yeah, agree 15:51:38 that's all from me today 15:51:44 #topic Open Discussion 15:51:54 I have one topic to discuss. 15:51:58 it is about error reporting 15:52:27 can I go? 15:53:15 sure 15:53:24 we've got 7 minutes more 15:53:25 horizon now catches an exception from back-end services and creates error messages in our code. 15:53:51 it hides the detail and we receive many bug reports on insufficient messages. 15:54:24 AFAIK, it is done from two reasons. 15:54:43 the one is to hide too *more detail* from back-end services like UUID 15:55:03 the other is for translation. we cannot translate messages from back-end services. 15:55:04 I'm OK to show exception messages from the APIs but not full tracebacks 15:55:29 however, I think detail information is more important than the above two reasons. 15:55:57 my proposal is to include exception strings to error messages. 15:56:13 e0ne: yeah, it is same as mine 15:56:48 I'm glad to be on the same page with the rest of the team 15:57:25 what in my mind is like this https://github.com/openstack/horizon/blob/master/openstack_dashboard/dashboards/project/networks/workflows.py#L499-L502 15:58:26 at least some leading messages can be translated :) 15:58:57 :) 15:59:39 i think we can do this gradually 16:00:24 that's all from me. 16:00:30 +1 16:00:45 thanks for the participation! see you next week 16:00:52 #endmeeting