*** ianychoi has joined #storyboard | 00:19 | |
*** diablo_rojo has quit IRC | 02:12 | |
*** mugsie has quit IRC | 03:22 | |
*** thrace has quit IRC | 03:22 | |
*** persia has quit IRC | 03:22 | |
*** mugsie has joined #storyboard | 03:38 | |
*** thrace has joined #storyboard | 03:38 | |
*** persia has joined #storyboard | 03:38 | |
*** irclogbot_2 has quit IRC | 03:41 | |
*** irclogbot_1 has joined #storyboard | 03:43 | |
*** udesale has joined #storyboard | 03:53 | |
*** whoami-rajat has joined #storyboard | 04:47 | |
*** udesale has quit IRC | 06:34 | |
*** udesale has joined #storyboard | 06:37 | |
*** tosky has joined #storyboard | 07:15 | |
*** dtantsur|afk is now known as dtantsur | 07:44 | |
*** udesale has quit IRC | 07:56 | |
*** jpich has joined #storyboard | 07:57 | |
*** zbr|ooo is now known as zbr|ruck | 08:01 | |
*** jtomasek has joined #storyboard | 08:38 | |
*** udesale has joined #storyboard | 08:55 | |
*** udesale has quit IRC | 10:35 | |
*** udesale has joined #storyboard | 10:38 | |
*** jpich has quit IRC | 10:50 | |
*** jpich has joined #storyboard | 10:51 | |
*** udesale has quit IRC | 11:48 | |
Zara | so SUSE has hackweek coming up, where you can work on whatever you want for a week as long as you're learning a new skill. vague plan is to hook up some kind of robot to storyboard that does something pointless when some conditions are met: THE STORYBOT. I have an arduino starter kit, a raspberry pi and a lego mindstorms set lying around. | 13:31 |
---|---|---|
Zara | so let's put storyboard on wheels | 13:31 |
Zara | ideas, recommendations etc welcome; not done any IoT before | 13:33 |
*** whoami-rajat has quit IRC | 13:57 | |
clarkb | I'm going to have to watch some storybots on netflix with the kids for research/ideas. Maybe have a robot flip a mechanical counter on say story closure | 14:04 |
Zara | :D | 14:05 |
fungi | most of the good ideas i can think of unfortunately require an event stream or webhook mechanism implemented in sb | 14:06 |
fungi | which i believe is something we want, but has lagged in priority behind more pressing issues | 14:06 |
SotK | simply subscribe to all projects and trigger things on receiving emails | 14:09 |
Zara | or I leave storybot with astonishing lag, so that it polls eg: 2x a day and if anything has changed in the instance, it makes a sound. a longwinded method to sound the alarm in the case of extended downtime | 14:17 |
fungi | oh, good suggestion SotK | 14:17 |
Zara | well, if it made a sound when nothing had changed | 14:17 |
fungi | in that case, there's always the classic foam dart launcher robot to get your attention each time a new story is filed | 14:17 |
Zara | :D | 14:19 |
SotK | (on a serious note, utilising the existing events we publish to rabbitmq to provide a neat way to connect to a websocket to consume them would be amazing) | 14:20 |
Zara | haha, but I'm hoping to make something thoroughly useless! :D guess we'll see what happens. | 14:23 |
SotK | heh, I just meant in general rather than trying to persuade you to do it, so don't feel any pressure to do so please :) | 14:39 |
Zara | dw :P just prepare for the world of the task-tracking battlerobots, in which priority is determined by combat. | 14:47 |
Zara | looks like my hackweek has morphed into an unreadable scifi novel before it even started | 14:47 |
Zara | (actually I was musing this morning about writing some story where the comedy twist at the end is that the characters are nodes in a HA cluster and there's some STONITH afoot) | 14:49 |
Zara | then I decided I probably needed to get a life | 14:50 |
fungi | SotK: yeah, assuming there's nothing sensitive on the message queue, that would make the most sense. we could even pick a standard protocol like mqtt (which is often exposed via websockets) | 15:01 |
fungi | https://www.rabbitmq.com/mqtt.html | 15:02 |
fungi | might not need any hacking *in* storyboard to make that happen | 15:03 |
fungi | and we've already got some experience with mqtt (including over websockets) https://docs.openstack.org/infra/system-config/firehose.html | 15:04 |
SotK | the downside to that approach is that most of the useful things to do with it are more interesting with support for private things | 15:06 |
fungi | yeah, probably so, and that means authentication and acls | 15:07 |
fungi | presumably we could mark events as "sensitive" and then an initial implementation would just filter those from the mqtt emitter | 15:08 |
fungi | but care would need to be taken to avoid bugs which leaked things we intended to be sensitive but botched the filter for or forgot to mark | 15:08 |
fungi | we could take a whitelist approach and tag all messages which we know *aren't* sensitive as public, and then only emit those via mqtt | 15:09 |
fungi | anyway, handwavey brainstorming | 15:09 |
persia | Is it not safe to use the "private" flag as a filter for "sensitive"? | 15:43 |
fungi | that i don't know, but it might be | 15:45 |
* SotK would expect it to be | 15:45 | |
fungi | depends on what exactly is actually going across the queue | 15:45 |
fungi | for a terrible example, if we needed to put events about session tokens across the queue, would those necessarily be marked as private? would we remember to do that for sure? | 15:46 |
fungi | events about stories and tasks are probably easy to solve that for, but it's not necessarily the case that's all which will ever be useful for the inter-process message queue | 15:47 |
*** jamesmcarthur has joined #storyboard | 15:56 | |
*** dtantsur is now known as dtantsur|afk | 16:26 | |
*** jpich has quit IRC | 16:29 | |
*** jamesmcarthur_ has joined #storyboard | 18:17 | |
*** jamesmcarthur has quit IRC | 18:21 | |
*** jamesmcarthur_ has quit IRC | 18:24 | |
*** jamesmcarthur has joined #storyboard | 19:02 | |
*** diablo_rojo has joined #storyboard | 19:33 | |
*** jamesmcarthur has quit IRC | 19:41 | |
*** jamesmcarthur has joined #storyboard | 19:42 | |
*** jamesmcarthur has quit IRC | 19:44 | |
*** jamesmcarthur has joined #storyboard | 19:44 | |
*** yoctozepto has joined #storyboard | 19:59 | |
fungi | anybody able to guess what could cause sb to send duplicate e-mail notifications to the same user when creating a story or adding a task? | 19:59 |
fungi | yoctozepto commented in #openstack-infra about having experienced that | 20:00 |
*** jamesmcarthur has quit IRC | 20:00 | |
yoctozepto | clarkb also affected | 20:00 |
fungi | indeed, clarkb do you happen to know if this is a recent change in behavior? | 20:01 |
fungi | could be a regression in one of the recent changes we merged, i suppose | 20:01 |
yoctozepto | I am not subscribing the project btw | 20:01 |
clarkb | I don't but can ask gmail | 20:01 |
yoctozepto | I have no history, that was my first post on storyboard | 20:02 |
clarkb | Looks like it may have started yesterday | 20:02 |
yoctozepto | ook | 20:02 |
clarkb | June 13 appears to be first case of it judging by quick scan of emails it sent | 20:02 |
fungi | yoctozepto: anyway, to answer your original question from the other channel, at a minimum i would consider it a bug, whether it's a long-standing one or not i couldn't tell you | 20:02 |
fungi | sounds like it could be a very recent regression though | 20:02 |
yoctozepto | thanks, fungi | 20:02 |
*** jamesmcarthur has joined #storyboard | 20:04 | |
*** jamesmcarthur has quit IRC | 20:05 | |
SotK | that sounds like a strange one | 20:27 |
SotK | I wonder if somehow two workers managed to process the events | 20:29 |
SotK | I actually saw a similar bug with the mqtt patches that mtreinish sent a while back, where each event would show up a bunch of times for some reason | 20:29 |
SotK | I figured it was a difference in mqtt vs rabbitmq, but maybe not | 20:30 |
* fungi checks for an excess of worker processes | 20:34 | |
fungi | nope, parent "storyboard-worker-daemon --worker-count 5" process and 5 children, as you would expect | 20:36 |
SotK | so I have a non-duplicated email from yesterday at 16:54 UTC | 20:40 |
SotK | and a duplicated one from 17:55 UTC | 20:41 |
fungi | that does indeed suggest something changed | 20:41 |
fungi | clarkb: are you able to help narrow down that window further? | 20:41 |
* clarkb looks | 20:42 | |
fungi | worst case i can try doing some mta log analysis, but that is not compatible with my present lawn care activities | 20:42 |
clarkb | no dup at 9:54am Pacific dup at 10:55am pacific yesterday | 20:43 |
clarkb | probably the same emails SotK got :) | 20:43 |
SotK | the only thing which merged in that timeslot was https://review.opendev.org/#/c/661663/ | 20:44 |
SotK | http://zuul.openstack.org/build/87e64ea9b90749d1b31f10f0cf1531e2 is the post job for that change | 20:45 |
*** zbr|ruck has quit IRC | 21:13 | |
*** zbr has joined #storyboard | 21:30 | |
*** diablo_rojo has quit IRC | 22:32 | |
*** tosky has quit IRC | 22:48 | |
*** jtomasek has quit IRC | 23:19 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!