15:59:44 #startmeeting api sig 15:59:45 Meeting started Thu Jan 11 15:59:44 2018 UTC and is due to finish in 60 minutes. The chair is elmiko. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:59:46 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:59:48 The meeting name has been set to 'api_sig' 15:59:53 oooh 16 seconds early 16:00:11 #chair cdent elmiko edleafe dtantsur 16:00:12 Current chairs: cdent dtantsur edleafe elmiko 16:00:19 #link https://wiki.openstack.org/wiki/Meetings/API-SIG#Agenda 16:00:27 o/ 16:00:35 #topic previous meeting action items 16:00:36 * edleafe comes out of hiding 16:00:57 #link http://eavesdrop.openstack.org/meetings/api_sig/2018/ 16:01:32 I did the writing for the annual report and was told "it's pefect!" 16:01:35 so I guess that went well 16:01:43 \o/ 16:01:50 cdent++ 16:02:22 huzzah! 16:02:33 :) 16:02:36 #topic open mic and ongoing or new biz 16:02:37 except that the quoted person spelt perfect perfectly 16:02:53 perfectly! 16:03:18 ok, so i see a topic of how do we encourage change without becoming the api police 16:03:22 any thoughts? 16:03:46 related to that, monty has proposed a couple of openstack-wide goals related to api stuff, including the recent pagination stuff 16:04:05 #link https://wiki.openstack.org/wiki/Meetings/API-SIG#Agenda 16:04:08 ooopes 16:04:11 #uncod 16:04:14 #undo 16:04:15 Removing item from minutes: #link https://wiki.openstack.org/wiki/Meetings/API-SIG#Agenda 16:04:17 * cdent is standing 16:04:19 a starting guide and a template for new projects? 16:04:22 #link https://review.openstack.org/#/c/532627/ 16:04:26 and openstack-wide goals indeed 16:04:30 * edleafe thinks cdent is wobbling 16:05:13 dtantsur: by template do you mean stub code or something less specific? 16:05:30 cdent: similar to cookiecutter 16:05:44 dtantsur: well, some of this was in relation to how we encourage projects to use the pagination rfc stuff, like should we open bugs or ? 16:05:48 something, preparing the python code and one CRUD resource with eerything set up 16:05:58 major versions, microversions, discovery... 16:06:13 dtantsur: I'd be pretty leary of code, but a cookbook on what the code can do would be useful 16:06:18 ooh, are you talking about a project template? 16:06:19 leery 16:06:32 * dtantsur opens google translate 16:06:47 cookbook ++, we went around the code template idea once before and it wasn't generally deemed worth it 16:06:52 too many options, too many variables 16:07:15 and too many opinions 16:07:30 ++ 16:07:35 actually, we may want to get into the business of chosing the recommended framework for new projects 16:07:45 and making sure it has enough libs for common stuff 16:07:47 only cdent's opinion counts, thoguh 16:07:49 though 16:07:53 agreed 16:08:10 dtantsur: how do you mean? 16:08:27 dtantsur: what would be an example? 16:08:51 edleafe: e.g. "use Flask; use for Restful API, for microversions, plug keystonemiddleware this way" 16:09:38 dtantsur: that sounds more like the stuff that the (now defunct) Architecture WG was focusing on 16:10:03 to me it sounds like something we could help with, even though it does grow our missing a tiny bit 16:10:35 edleafe: perhaps api-wg would not address such things, but the api-sig might 16:10:35 otherwise we have a lof of recommendations, but no practical help in implementing them.. 16:10:53 it can mean some understandable resistance 16:10:57 s/mean/meet/ WUT? 16:11:01 such advice would be something other than "guidance" 16:11:43 yeah 16:13:07 there's nothing new on api-schema is there? 16:13:18 i had always thought that adding some code projects would help drive some traffic to the sig(ex. wg), but it never seemed to gain much traction 16:14:04 cdent: i don't think so, but i also don't know for sure 16:14:22 I think I just left it on the agenda as a reminder 16:14:36 so, do we have anything to record about encouragin projects to adopt the guidelines for things like pagination? 16:15:04 last time we talked about opening bugs on projects that might need it, is this something we would pursue? 16:15:20 do you all agree that it is okay for them to sometimes be an openstack-wide goal? 16:15:33 or is that too stick-ish? 16:15:46 sometimes - yes 16:15:47 elmiko: there's always this: https://goo.gl/images/KkRL2o 16:15:49 man, that seems so heacy 16:15:52 heavy* 16:16:08 edleafe: LOL, not quite my style but yeah 16:16:50 cdent: aren't goals supposed to be complete-able in one cycle? 16:17:02 cdent: my question to that would be, if we say something is an openstack-wide goal (eg pagination either in headers or body), then how do we get buy-in from the community for our pronouncements? 16:17:23 edleafe: yeah, thus why the pagination thing is just one thing 16:17:48 elmiko: the goal has to be approved by the tc and the tc won't approve it if there isn't perceived buy in 16:17:57 and if there isn't a champion to drive it 16:18:01 and would we need a separate process for declaring something a goal like that? 16:18:09 cdent: yeah, that's what I thought 16:18:33 elmiko: we don't really do the declaring (unless we feel like it) see 16:18:37 so, i'm ok with the openstack-wide goal if we have a higher body to help us (ie the tc) 16:18:46 #link https://governance.openstack.org/tc/goals/index.html 16:18:57 So saying "adding/correcting pagination links" is doable, but "conforming to the API guidelines" would not be 16:19:39 i had not seen that page, given that i am ok with going forward on the goals 16:19:39 ? 16:19:56 we miss you in openstack elmiko 16:20:09 <3 i miss y'all too =) 16:21:25 ok, so for this week is it fair to say that we agree the sig will works towards identifying pieces of the guidelines with high impact that should be promoted to openstack-wide goals? 16:21:55 there's only two goal (usually) per cycle, so this is not a solution, simply a tool 16:22:04 ah, gotcha 16:22:05 but it would be safe to say that we seem okay to use the tool 16:22:21 still though, getting pagination fixed across the ecosystem in a single cycle would be a big accomplishment 16:22:32 yes 16:22:47 ok then, moving along 16:22:54 anyone wanna talk about health check apis? 16:22:57 #link https://etherpad.openstack.org/p/self-healing-rocky-forum 16:23:04 #link https://etherpad.openstack.org/p/sydney-cloud-native-partii 16:23:19 (i'm not sure who added this) 16:23:47 I'm not sure from these links what we should talk about right now 16:24:05 ack 16:24:14 sorry, that was me. 16:24:19 i wasn't sure if we had an advocate in attendence 16:24:51 there's work in progress to establish health check tools (things that make it easy to kill contiainers when not healthy) 16:25:24 the api-sig has been invited to help define the api for the for that, so that the implementation is consistent across services 16:25:29 things like the same well known url 16:25:34 there's an oslo-spec in progress 16:25:49 #link https://review.openstack.org/#/c/531456/ 16:26:01 i think it is suffering from internet over-engineering task force 16:26:05 but lays out some of the goals 16:27:09 * elmiko chuckles 16:27:18 i had not heard ioetf XD 16:27:27 heh 16:27:42 I'd say this document should first appear in api-sig indeed 16:27:59 the oslo-spec is mostly about the backend 16:28:55 but as the api-sig, we're under no obligation to only attend to documents in our domain, being active on that spec is desired and invited 16:29:34 seems like there could be two parallel efforts: the oslo one to implement, and the api-sig to review and create the guidelines as needed 16:29:55 edleafe++ 16:30:23 * cdent nods 16:30:44 yeah.. thought I'd still prefer the api-sig one to show up first 16:31:00 too late? :) 16:31:05 there are interesting things from our perspective that seem missing: like how to manage microversions? 16:31:08 heh 16:31:24 is it going to have separate microversions? the same as project? none? 16:32:34 that, too, sounds like ioetf 16:32:44 this thing should answer yes, no, and maybe maybe 16:32:46 I guess I would have to know more of the health check mechanisms to consider the need for versioning 16:33:18 reading the specs and the etherpads will help to illuminate some of the thinking, especially the comments on the spec will identify some of the different ideas 16:33:28 * dtantsur puts on his review list 16:33:31 there are some preconceptions which could probably be adjusted usefully 16:34:27 aside from the guidance about which libs to use and code stuff, i think this would really be nice to land as a guideline 16:34:36 *assuming* that a concensus can be achieved 16:35:07 * elmiko is skimming over the review 16:35:14 * cdent nods 16:37:33 i realize we might not be as quick to react as the folks driving that review though, so i'm curious if there is a way for us to signal this intention to those folks and get their advice 16:37:49 comment on the review? 16:37:52 just say so on the review 16:38:03 that sounds way too reasonable 16:38:30 #action elmiko add thoughts to health check review about becoming an api-sig guideline 16:38:32 ok, then show up at their homes and bang on their doors 16:38:35 sound good? 16:38:38 LOL 16:38:51 * edleafe likes being unreasonable 16:38:57 ++ 16:39:27 take that gun from earlier 16:39:28 hehe 16:39:37 cdent: fair enough lol 16:40:10 moving on... 16:40:12 #topic guidelines 16:40:19 #link https://review.openstack.org/#/q/status:open+project:openstack/api-wg,n, 16:40:25 #link https://review.openstack.org/#/q/status:open+project:openstack/api-sig,n,z 16:40:38 there was that addition from mordred, anything else? 16:40:40 #link https://review.openstack.org/#/c/532814/ :) 16:40:56 ooh! 16:41:07 dtantsur's new thing is vera nice 16:41:31 If dtantsur likes https://review.openstack.org/#/c/531914/ we should probably freeze it? 16:41:42 agreed 16:41:58 cdent: I haven't checked it yet, but you can freeze it. I will let you know if I really object :) 16:42:21 freeze it, baby! 16:42:41 frozen 16:43:05 ok, and we need to do some reviews on dtantsur's 16:43:09 well, i do at least 16:43:28 yeah, I need to review, too 16:43:54 MOAR REVIEWS \o/ 16:43:59 \o/ 16:44:02 #topic bug review 16:44:09 #link https://bugs.launchpad.net/openstack-api-wg 16:44:15 #link https://bugs.launchpad.net/openstack-api-sig 16:44:23 anything new here? 16:44:47 no sir 16:45:17 #topic weekly newsletter 16:45:21 volunteers? 16:45:38 it's been a while since I have, so... me? 16:45:52 i'm cool with that, i have a preso after this meeting anyways 16:45:55 I'll go sit somewhere, so I can actually type 16:46:00 * edleafe pats cdent on the back 16:46:06 for completeness 16:46:06 I know, right? 16:46:07 #link https://etherpad.openstack.org/p/api-sig-newsletter 16:46:25 alrighty then, any last comments? 16:46:35 I'll ping when I have something to review 16:46:44 ++ 16:46:57 ok, we get back 14 minutes. thanks all! =) 16:47:03 #endmeeting