Service Ownership @Slack
summary
Slack moves from a centralized ops team to dev teams that own the full lifecycle, with embedded SREs mentoring teams, service-level criteria, and one afternoon when engineers turned off low-level alerts to force ownership. Numbers matter: Slack scales from ~100 servers to ~15,000, 10M daily users, and devs deploy over a hundred times a day. Result: dozens of team-specific on-call rotations, clearer alerting, measurable reductions in noisy pages, and a still-evolving post-incident process.
AI-written summary · about
chapters (11)
- 0:00Start
- 5:32Kaizen Continuous Improvement
- 8:02Executive dedication to learning
- 11:17Who should be responsible for the management, monitoring and operation of a production application?
- 17:34Postmortems weren't a great place for learning
- 18:41Centralized Operations Service Ownership
- 19:04High Trust Learning Organization
- 20:54Plan: Send these paging alerts to the development teams
- 23:27Test with the users
- 25:22Empowered Continuous Improvement
- 28:48Copy the questions Not the answers
related talks
26mService Ownership: DevOps for Salesforce
20mOptimizing Delivery with SlackOps
39mThe Life of On-Call, how SRE is changing DevOps
31mScaling SRE at MailChimp: How We Did It
18mSLICK: Driving SLO Culture
44mFull Service Ownership & The Lifecycle of a Service
39mSRE or Devops - I am an Architect - Why Do I Care?
26mIncreasing the collaboration between SREs and Developers