Wikimedia Release Engineering Team/Quarterly review, November 2013/Notes

Noles from the Q2 Release Engineering + QA Quarterly Review

Attendees: Jeff, Greg, Erik, Chris S, RobLa, Chris M, Zeljko, Andre, Ori, Sam, Dan

Last quarter's goals

Things that we weren't able to get to:

Better align QA
- Increase capacity through community training for browser tests
Improve our deployment process
- automate as much as possible
- improve tooling
Take Beta Cluster to next level
- automated test for the API
- feed experiences from Beta to production

Beta cluster certs - identifying a subset of certs that will get us the most bang for the buck

git-deploy: need to work out ownership

Sam is slightly leaning towards git-deploy even though it could be a long haul to get to feature parity to all the hacks ontop of scap

ACTION: Need to document the feature requirements of sartoris/etc - possible task for Bryan Davis after scholorship app

ACTION: clarify priority of work with antoine re vagrant spin ups for Jenkins builds

Monitoring:

http://ganglia.wmflabs.org <- It's up, but Antoine could use help from ops on this
Antoine would like to make tests against Vagrant instances triggered by Jenkins as a priority

Questions:

Q: who is working on Vagrant support?
- A: mostly distributed (e.g. Adam, Yuvi, etc)
Tomasz: what is being done to distribute work to other teams?
- Greg: I think that's what our first bullet point was trying to get at
- Chris: we have a great model working with the Language team

Resourcing:

Ops support: not in the next 2 months, realistically
Deployment and tooling related:
- resourcing for tooling dev
- begin with a sprint to kick off dev tooling with aaron/ori (arch) and bryan (dev)?
- delegate from Aaron: FileBackend
- maybe easier to do limited bursts of support with an explicit backlog of issues for a specific component (eg ResourceLoader)