| *** saju_m has joined #storyboard | 06:47 | |
| *** openstack has quit IRC | 08:21 | |
| *** openstack has joined #storyboard | 08:30 | |
| ruhe | ttx: do you think it's going to become mainstream in OpenStack? | 08:30 |
|---|---|---|
| ruhe | will other projects adopt this approach? | 08:30 |
| *** openstackstatus has joined #storyboard | 08:30 | |
| ttx | ruhe: haven't read the thread yet | 08:33 |
| ttx | ruhe: it's compatible with storyboard though | 08:33 |
| ttx | ruhe: we'd just have a task that affects the nova-specs repo | 08:34 |
| *** jcoufal has joined #storyboard | 08:36 | |
| ruhe | ttx: i just imagine a tight integration with this - once spec patch is approved, corresponding story also moves to approved state. in other words story state would automatically reflect the state of spec patch | 08:47 |
| SergeyLukjanov | ruhe, IMO if we have good enough comments and ACL systems in storyboard than there will be no need for such approach | 08:48 |
| ruhe | SergeyLukjanov: you mean - let drivers team to update story state manually? | 08:49 |
| SergeyLukjanov | ruhe, I mean that profit from the CR for BP approach will be less meaningful | 08:50 |
| ttx | ruhe: or not having "story state" at all | 08:51 |
| ruhe | SergeyLukjanov: i see your point. yes, with the right design StoryBoard will let people do what they're doing now with BP/gerrit | 08:52 |
| *** jcoufal has quit IRC | 09:20 | |
| *** persia has joined #storyboard | 09:32 | |
| *** persia has quit IRC | 09:32 | |
| *** persia has joined #storyboard | 09:32 | |
| *** saju_m has quit IRC | 10:17 | |
| *** tteggel has quit IRC | 10:18 | |
| *** tteggel has joined #storyboard | 10:18 | |
| *** saju_m has joined #storyboard | 10:32 | |
| *** jcoufal has joined #storyboard | 11:09 | |
| *** openstackgerrit has quit IRC | 11:10 | |
| *** openstackgerrit has joined #storyboard | 11:10 | |
| *** saju_m has quit IRC | 11:15 | |
| *** saju_m has joined #storyboard | 11:36 | |
| ruhe | anyone wants to comment on https://review.openstack.org/80836 ? | 13:12 |
| *** hashar has joined #storyboard | 13:19 | |
| *** mfer has joined #storyboard | 13:19 | |
| *** jcoufal has quit IRC | 14:03 | |
| *** jcoufal has joined #storyboard | 14:04 | |
| krotscheck | ruhe: I'll take a look | 14:39 |
| ruhe | krotscheck: danke schön | 14:40 |
| krotscheck | ruhe: Does python or pep8 enforce a "first statement of an override must be super()" rule? | 14:50 |
| krotscheck | I'm curious on whether it's possible for that addCleanup call to somehow end up in a weird order-of-operations error. | 14:51 |
| *** che-arne has joined #storyboard | 14:54 | |
| krotscheck | You know, even if it did, that'll fail almost instantly. | 14:56 |
| ruhe | krotscheck: it doesn't, but it's a very common practice to put it as a first statement | 15:03 |
| *** saju_m has quit IRC | 15:03 | |
| *** saju_m has joined #storyboard | 15:20 | |
| ruhe | krotscheck: also i spoke with our folks who drive oslo.db. they said that approach we're taking in our DB tests is the right approach | 15:20 |
| ruhe | also they're working on implementing parallel tests on real DBs | 15:20 |
| ruhe | so, we just need to wait :) | 15:20 |
| ruhe | krotscheck: NikitaKonovalov: could you please checkout my patch and see if it works correctly on your local machines? | 15:23 |
| NikitaKonovalov | ruhe: checking ... | 15:25 |
| krotscheck | ruhe: Willdo | 15:25 |
| krotscheck | ruhe: Had to blow away the *.pyc files before it worked, but py27 seems to pass. Checking the others now | 15:27 |
| *** david-lyle has joined #storyboard | 15:28 | |
| krotscheck | pypy hates me | 15:45 |
| *** miqui has joined #storyboard | 15:47 | |
| openstackgerrit | Ruslan Kamaldinov proposed a change to openstack-infra/storyboard: Fix DB migrations in unit tests https://review.openstack.org/80836 | 15:50 |
| krotscheck | Thanks for fixing that | 15:50 |
| mordred | ++ | 15:53 |
| krotscheck | mordred: I need your database brain | 15:58 |
| krotscheck | mordred: Given what we've currently built, and where it's going, we're likely to have the use case of stackalytics and/or ttx writing overview reports of activity. I'm worried that reports like that will generate ad-hoc filters and queries that will block the database and compromise storyboard. Do you think that's a likely scenario? | 16:00 |
| krotscheck | In a nutshell: I don't want jenkins/zuul to start throwing errors when they try to update tickets, because someone's using the API to load the entire database contents so they can have a pretty chart. | 16:01 |
| openstackgerrit | A change was merged to openstack-infra/storyboard: Fix DB migrations in unit tests https://review.openstack.org/80836 | 16:12 |
| cody-somerville | krotscheck: Do you think there is something architecturally wrong with what we've built that would cause that? | 16:14 |
| cody-somerville | krotscheck: Unless the api calls involve locks, it shouldn't be too much of a problem and there are known solutions to scaling. | 16:15 |
| *** saju_m has quit IRC | 16:15 | |
| krotscheck | cody-somerville: Architecturally speaking? Yes: We're using an ORM. | 16:16 |
| cody-somerville | krotscheck: Lot's of websites use ORMs. It certainly has a level of overhead and can sometimes result in some inefficient queries but we can scale that out horizontally. | 16:18 |
| krotscheck | cody-somerville: I'm familiar with quite a few scaling approaches as well - I just think that it might be better (at least in the short run) to provide a replicated, read only database where ttx can write his own queries. That way he can join all he wants. | 16:18 |
| cody-somerville | krotscheck: I imagine we can do that if we find we need to. I believe sqlalchemy has a 'router' type system similar to Django that lets you route read-only requests to your read-only replicas. | 16:19 |
| krotscheck | cody-somerville: Well, ok, let me give you some background - I've been burned because a BI's reporting query has taken down a live e-commerce site. | 16:19 |
| krotscheck | cody-somerville: So whenever someone says "Hey I want to run reports on this" my immediate response is "oh hell no" | 16:20 |
| *** jcoufal has quit IRC | 16:24 | |
| cody-somerville | haha @ BI reporting taking down a e-commerce site. that's gotta hurt. | 16:24 |
| krotscheck | ....yeeeaaaah. | 16:26 |
| ttx | krotscheck: i epect I won't be the only one hitting the API for basic reports though | 16:30 |
| ttx | expect, even | 16:30 |
| *** hashar has quit IRC | 16:36 | |
| krotscheck | ttx: Exactly. For now I'd be happy if we can get some notification of the request/response times, so we keep tabs on what queries take forever. | 16:44 |
| ttx | krotscheck: back from vacation, will help with storyboard as soon as I'm done with the backlog | 16:45 |
| krotscheck | ttx: ruhe just figured out the sqlite nonsense, you should be able to rebase/fix your migration | 16:47 |
| krotscheck | ttx: We did rename the migrations though, fair warning. | 16:48 |
| ttx | krotscheck: I should be away more often | 16:48 |
| krotscheck | NikitaKonovalov, ruhe: jeblair's requested that we get story & task comments working next. How do your schedules look regarding the work items we discussed on thursday? | 16:51 |
| krotscheck | mordred: does https://review.openstack.org/#/c/80836/ supercede https://review.openstack.org/#/c/79208/ ? | 17:01 |
| NikitaKonovalov | krotscheck: we've got a db model for story comments, so task commnets should be mostly the same | 17:02 |
| krotscheck | NikitaKonovalov: Let's start with stories then. | 17:02 |
| NikitaKonovalov | then db_api layer is pretty easy to do | 17:02 |
| krotscheck | Yeah, I suspect the UI is going to be tha hard part :) | 17:03 |
| NikitaKonovalov | an the only remaining thing is a controller | 17:03 |
| krotscheck | NikitaKonovalov: Something about keeping comments in different tables for different asset types feels like duplicating code to me :/ | 17:04 |
| NikitaKonovalov | we can do that with inheritance and partial overriding I guess | 17:05 |
| krotscheck | NikitaKonovalov: Yeah, thankfully we don't have to do a fulltext search on comments and return different resource types. | 17:06 |
| krotscheck | NikitaKonovalov: Where should the comments endpoint live? /story/ID/comments/ID? | 17:07 |
| krotscheck | NikitaKonovalov: Treat it as a subresource of stories? Or as its own resource? | 17:07 |
| krotscheck | i.e. /story/ID/comments/ID vs /comments/ID | 17:07 |
| NikitaKonovalov | we can have two endpoints /story/id/commnets and /task/id/comments | 17:10 |
| krotscheck | NikitaKonovalov: Yeah, I'm just debating whether it makes sense to treat comments as an independent resource or as a subresource of the thing it's related to. | 17:13 |
| * krotscheck can see both sides of the argument. | 17:13 | |
| krotscheck | ruhe, cody-somerville, ttx: Thoughts on API endpoint for comments? ^^ | 17:14 |
| ttx | krotscheck: I don't think we want comments for tasks. only for stories. | 17:16 |
| ttx | krotscheck: I would place it under stories, but hen I'm not that RESTful | 17:17 |
| ttx | /story/ID/comments/ID | 17:17 |
| krotscheck | ttx: I don't think that's an issue with RESTfulness, since we're describing an explicit subresource. | 17:17 |
| krotscheck | Alright, if no comments for tasks, then subresource it is. | 17:18 |
| ttx | krotscheck: you are the RESTmaster | 17:18 |
| krotscheck | *snort* | 17:18 |
| * ttx is back to no unread mails ! | 17:18 | |
| * ttx was back to no unread mails. | 17:18 | |
| krotscheck | If I was the rest master, I'd demand that foreign keys be replaced with fully qualified resource URI's. | 17:18 |
| krotscheck | But that's really just being pedantic. | 17:22 |
| krotscheck | And pedantry isn't funny | 17:22 |
| *** hashar has joined #storyboard | 19:48 | |
| ruhe | krotscheck: re schedules - i have more work on this week than i can actually do :), so if you leave me something to hack on i'll do that at the weekend. but i guess NikitaKonovalov will be able to cover REST API side | 19:53 |
| ruhe | krotscheck: re API endpoint for comments i agree with ttx | 19:54 |
| ruhe | krotscheck: re performance - we always can run (i hope so) several StoryBoard instances behind a load balancer | 19:54 |
| ruhe | and we will just let mordred to figure out DB scaling issues :) | 19:55 |
| krotscheck | Got it | 20:06 |
| *** mfer has quit IRC | 20:59 | |
| *** hashar is now known as hasharMeeting | 21:11 | |
| * krotscheck found a bug in angular. | 21:52 | |
| *** david-lyle has quit IRC | 22:04 | |
| openstackgerrit | Michael Krotscheck proposed a change to openstack-infra/storyboard-webclient: Added user preference handling https://review.openstack.org/81100 | 22:30 |
| *** openstackgerrit has quit IRC | 22:39 | |
| *** openstackgerrit has joined #storyboard | 22:40 | |
| *** hasharMeeting has quit IRC | 22:44 | |
| *** miqui_ has joined #storyboard | 23:42 | |
| *** miqui has quit IRC | 23:43 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!