DevOps Ten Years After: Review of a Failure with John Allspaw and Paul Hammond
summary
Ten years on, the claim that teams ran continuous deployment and even developers deploy their own code is defended with concrete shifts: configuration management from CFEngine → Terraform, IRC bots to Slack bots, and feature flags used as hedges against uncertainty. The surprising operational detail is that deployment rights and readable ops repos are the real capability, not cadence alone. Discussion covers buy vs build tradeoffs, page-rotation burnout, and measuring whether observability tools and dashboards are actually used.
AI-written summary · about
chapters (16)
- 0:00Introductions and Context
- 2:58The Original Velocity Talk
- 5:39The Shift to Modern Operations
- 9:11Industry Evolution and Impact
- 12:18Emergence of DevOps Culture
- 15:35The Silver Bullet Cycle
- 19:07DevOps and Business Value
- 22:19Hedges Against Uncertainty
- 26:50Build vs Buy Trade-offs
- 29:36Financial Engineering
- 32:21Visibility and Tool Utility
- 34:47Progress in Cultural Norms
- 37:40Transparency and Code Review
- 41:11Learning and Collaboration
- 43:16Future Challenges: Burnout
- 45:27The Path Forward
related talks
44mDevops - Almost 10 Years - What A Strange Long Trip It's Been (Keynote)
32mLooking Back At 13 Years Of DevOps
51mFive Years of DevOps: Where Are We Now?
27m10 years of #devops, but what did we really learn?
49mTen Years of the DevOps Movement - A Retrospective and Outlook
56m14 years of #devops, but what did we really learn
40mDistilled Abnormal DevOps Iterations
19m10 years of DevOps - Where Are We Today?