09:00:01 #startmeeting large_scale_sig 09:00:02 Meeting started Wed Jan 29 09:00:01 2020 UTC and is due to finish in 60 minutes. The chair is ttx. Information about MeetBot at http://wiki.debian.org/MeetBot. 09:00:03 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 09:00:05 The meeting name has been set to 'large_scale_sig' 09:00:07 #topic Rollcall 09:00:12 hello 09:00:14 o/ 09:00:19 Hi everyone! 09:00:24 hi 09:00:29 Agenda at: 09:00:33 #link https://etherpad.openstack.org/p/large-scale-sig-meeting 09:00:46 I reordered it a bit to give masahito time to join before we discuss oslo.metrics 09:01:21 amorin: around? 09:01:45 I must leave after 30 minutes, apologies 09:01:45 Hello 09:01:50 I am here 09:02:00 ok let's start then 09:02:01 Hello, Sorry for being absent for a long time. 09:02:05 Hi 09:02:11 #topic Progress on "Documenting large scale operations" goal 09:02:14 o/ 09:02:16 #link https://etherpad.openstack.org/p/large-scale-sig-documentation 09:02:25 We had the following action items from last meeting(s): 09:02:36 - Collect and add links to relevant articles around large scale openstack (if you find any) to the etherpad 09:02:46 We had some links filed from the Discovery initiative (experimentations on edge) 09:02:53 But otherwise not that much 09:03:04 But then maybe it's more of a long-term action, and we should just carry it over? 09:03:16 yes i think so 09:03:20 I think so. New things will turn up 09:03:31 It's more as we find them, to remember to add those to the etherpad :) 09:03:43 - oneswig to follow up with Scientific community to find such articles 09:03:57 oneswig: did you mention it on the Scientific SIG meeting ? 09:03:58 I added few links today. But this will be a working in progress 09:04:38 ttx: I did. We have meetings in two timezones and I raised it in both. I think people looked at the etherpad but I don't think anything more was added by them, unfortunately. 09:05:00 OK, I propose we have a standing item in meetings, to regularly review the list we collected so far... rather than an action item 09:05:12 However as above I'll keep looking for useful links from scientific sig discussions going forward. 09:05:15 since this will basically never be "done" 09:05:43 We can curate the list and categorize it on the etherpad itself 09:06:07 does that work for everyone? 09:06:18 yes 09:06:24 +1 09:06:32 The other action was around amorin's documenting configuration defaults for large scale 09:06:33 +1 09:06:41 amorin: How is that going ? 09:06:56 hey all 09:06:56 +1 09:07:11 I did not had enough time to work on this topic 09:07:52 That's fine. If you need help in specific areas just tell us 09:08:48 As I mentioned when we started this, it's ok to go slowly. Members of this group all have demanding day jobs :) The essential is to make regular progress 09:09:05 yup 09:09:23 Anything else on that topic? 09:09:59 I'll take that as a 'no' 09:10:00 #topic "Scaling within one cluster" goal 09:10:05 #link https://etherpad.openstack.org/p/large-scale-sig-cluster-scaling 09:10:14 We had the following action items from last meeting(s): 09:10:24 - all post short descriptions of what happens (what breaks first) when scaling up a single cluster 09:10:31 #link https://etherpad.openstack.org/p/scaling-stories 09:10:40 Nobody posted anything yet ^ 09:10:55 So... Is it that you did not find the time, or that you don't have anything to say, or somethign else? 09:11:13 (trying to see if it's still a good idea) 09:11:31 The investigation I have been doing has been on instrumentation, and there may be a scaling story in that. 09:11:41 personally, I don't have much to say because I do not operate a large scale cluster myself 09:11:55 Just I didn't have much free time to do it :-( 09:12:19 masahito: maybe just add a link to that presentation you did in Shanghai. I think that would be a great start :) 09:12:25 didn't find the time to document. For our case it was RabbitMQ, Nova-scheduler. Will add after the meeting. 09:12:58 ttx: ah, okay. I didn't think of it. I'll do that. 09:13:22 masahito: I enjoyed your presentation in Shanghai - please do :-) 09:13:23 My hope is that once we have a few things in there, we can ask again to the openstack-discuss list and get them to add to it 09:13:37 It's just difficult to seed the list initially 09:14:01 We can also make sure it's a topic discussed at future Ops Meetups 09:14:59 ok next... 09:15:02 - masahito to produce first draft for the oslo.metric blueprint 09:15:06 #link https://review.opendev.org/#/c/704733/ 09:15:47 yes. This is actually a draft. 09:15:55 I propose taht between now and next meeting, we all read the review the draft and comment as necessary 09:16:18 Nice work, will do. 09:16:27 so that we iterate toward a final proposal 09:16:47 Does that sound like a good idea ? 09:16:59 ok for me 09:17:08 I will review it 09:17:11 #action all to review oslo.metrics draft at https://review.opendev.org/#/c/704733/ 09:17:16 Feel free to add you comment 09:17:24 sorry, away from keyboard 09:17:45 - all learn more about golden signals concept 09:17:52 #link https://landing.google.com/sre/book.html 09:18:06 I'll admit I did not find the time to read that 09:18:25 me too 09:18:33 It's quite fluffy... the details could be condensed 09:19:00 I'll try to read it also 09:19:15 If possible I'd like to propose a kind of show and tell about some investigation we have been doing in this area 09:19:19 for the next meeting 09:19:19 Is there a specific chapter we should focus on ? 09:19:33 sure that would be perfect 09:19:54 probably this bit; https://landing.google.com/sre/sre-book/chapters/monitoring-distributed-systems/#xref_monitoring_golden-signals 09:20:12 ok, that sounds more reasonable to digest 09:20:41 Let's read that before next meeting as preparationfor your show-and-tell 09:20:50 One of my colleagues has been doing some good work on using telemetry from HAproxy in Monasca. 09:21:13 #action all learn about golden signals by reading https://landing.google.com/sre/sre-book/chapters/monitoring-distributed-systems/#xref_monitoring_golden-signals 09:21:15 There have been recent developments (in the last year or so) which really improve the data available about API service times 09:21:41 #action oneswig to prepare show-and-tell about some investigation they have been doing in this area 09:22:22 Will do. 09:22:28 oneswig: how much time will you need ? 09:22:55 Probably 15 minutes, depends on how much detail is wanted. 09:23:09 ok so it will fit the regular meeting ok 09:23:18 yes, no problem. 09:23:25 Anything else on that topic ? 09:24:17 not from em 09:24:18 #topic Other topics 09:24:19 me 09:24:23 Last meeting belmiro proposed a every-two-week cadence. 09:24:34 If that still works for everyone, I will officially add the Large Scale SIG meeting so that it appears on the official meetings agenda at http://eavesdrop.openstack.org/ 09:25:06 ok for me 09:25:13 #action ttx to add meeting to eavesdrop on a every-two-week cadence 09:25:17 #info next meeting Feb 12, 9utc 09:25:31 Is that ok for everyone ? 09:25:36 yes 09:25:38 it's ok for me 09:26:05 OK, is there anything else you wanted to discuss in this meeting? 09:26:21 anything else we should be doing between now and next meeting? 09:27:21 If not, I'll post a summary with all the action items we just discussed. Please try to complete them by next meeting :) 09:27:55 and close this one 09:28:02 oneswig: in time for your next meeting :) 09:28:07 nothing from my side 09:28:15 perfect, thanks :-) 09:28:35 Thanks everyone! Talk to you all again in two weeks! 09:28:41 thanks 09:28:51 Anyone from teh group present at FOSDEM this weekend? 09:29:01 Thanks everyone... 09:29:15 belmoreira, amorin oneswig maybe 09:29:43 I'm not going 09:29:43 sorry ttx, not this time for me at fosdem 09:30:05 OK, I'll represent :) 09:30:12 Thanks all 09:30:14 #endmeeting