Project Name: wmf-reading
Purpose: Shared testing and prototyping for Reading department teams
Wikitech Username of requestor: Jdlrobson
Admins: Jdlrobson, Jhernandez, phuedx, Bmansurov, BryanDavis
Rationalization:
Having a general purpose project for Reading will allow teams to create and manage VMs for acceptance testing and other general purpose collaboration tasks quickly and easily.
| Status | Subtype | Assigned | Task | ||
|---|---|---|---|---|---|
| Resolved | bd808 | T76375 [DO NOT USE] New Labs project requests (tracking) [superseded by #cloud-vps-project-requests] | |||
| Resolved | yuvipanda | T101325 Create labs project for Reading department |
@yuvipanda @scfc we want to experiment with a standalone instance to develop a bot. All we need is a box to run it where we can run selenium web driver and access Gerrit API endpoints. If there's a better place to do this fine.
See background in https://phabricator.wikimedia.org/T100293
in particular "We've talked with releng and we know that they are working on a generic solution, but it will take a lot of time. We need a temporary solution so we knowingly accept the cost of maintaining our own browser testing infrastructure but we feel that it's far outweighed by the increased confidence in our code. (Our products are very user exposed and we need a good integration tests story)."
How about creating a generic "readership team" labs project that we can host this and other experiments in? The MediaWiki-Core-Team had (has actually) a mediawiki-core-team project that we used for lots of different testing vms and other tools. With the umbrella project and Labs-Vagrant it is pretty easy to spin up a new VM that hosts a full stack wiki including central auth, VE and other nice goodies.
I think projects should be more specific and not be team based - that's something we've been trying to move away from in the past so we don't end up with projects with vague / no ownership.
If releng is fine with it, I can happily create a more narrow scoped project - 'mobile-smoketests' perhaps?
I just hate to see the need to make a ton of small, short lived projects both because it becomes difficult to know when to shut them down and it takes a while to spin a new one up. The Reading division owning the Reading project seems like a pretty clear provenance to me, but I will agree that one problem in a project managed by a larger group is finding out what the status of each instance is. If the general guidance of the Labs team is not to make these types of projects though I won't argue at all.
@bd808 I think the solution is to reduce the time it takes to spin a new one up, and we've been better about it now than with the previous SMW (praise be its name!) process (which averaged in weeks!). More guidance is probably needed there (Should file as subtask of T101659?), but ideally it should be under a day.
(I created mobile-smoketests when Jon pinged me on IRC, I'm going to consider this one done)
On hindsight a 'readership' project as @bd808 suggested would definitely have made more sense. Can you rename projects? I have a weird situation now where the instance and the project are pretty much the same but it's helping me get things done so that's good - so thank you!
I disagree - I think projects should be more spefific but just easier to create - it's totally ok to have an instance and project just be the same :)