Wikimedia Release Engineering Team/Ongoing maintenance

Let's define, and put rough numbers next to, the things that take up our time during the week or month. Things that take time away from working on new projects.

Number of hours per week is an imperfect/imprecise/sometimes flawed way of measuring this, but we should have some baseline/rough understanding of where our time is going.

Things

• Jenkins coddling - restarts/kicking
  • Antoine
  • # hrs/week
• Beta Cluster coddling - breakages from production puppet changes, kicking in other ways
  • Antoine, Bryan
  • # hrs/week (Bryan: 5 hrs/week)
• MediaWiki Release management - checkins with contractors, build/testing support, etc
  • Greg, Antoine, Chris
  • # hrs/week (Greg: 1 hr/week average)
• Deployments - just the update/deploy bits
  • Reedy
  • 2-3 hrs/week usually
• Deployments - SRE stuff
  • Reedy
  • 3 hrs/week usually
• Deployment response - aka fatal monitoring/bug reporting
  • Reedy
  • 2 hrs/week
• Shell requests/config gardening
  • 3ish hrs/week (maybe more, depending)
• Browser test pairing
  • Željko
  • 8 hrs/week
• diagnosing failing builds
  • Chris, Željko
  • # hrs/week (Željko 8 hrs/week)
• Reviewing code (Gerrit)
  • # hrs/week (Željko 8 hrs/week)
• more here