DevOps Counseling: 10 ways to help your Dev and Ops people love each other
summary
Most failures are social not technical: teams have great tools (CI, RPM packaging, Puppet/Chef, Nagios) yet run a website at 85% uptime because handoffs, incentives and silos block delivery. Concrete fixes: collapse an 11‑tab change request into a simple web form, embed ops in dev teams, share metrics for A/B rollback, give trusted engineers root on servers or paired access, run packaging special interest groups, and use rituals like the fail cake to diffuse blame and teach quick recovery.
AI-written summary · about
chapters (35)
- 0:00Intro
- 8:28Different teams
- 9:42Production
- 10:43Agile Development
- 13:27Sitting Together
- 16:21Open Channels
- 18:48IRC Terminal
- 19:44Knowledge Sharing
- 22:02Irrelevant Learning
- 23:05Conways Law
- 27:45Similar reporting
- 29:22Cost of communicating
- 30:11Cost of gates
- 32:52Value stream mapping
- 33:55Incentives
- 34:07Who doesnt like getting paid
- 34:58Do it fairly
- 35:40Dont get games
- 37:42Celebrate
- 40:58Learn
- 41:40Celebrate small failures
- 41:52Fail cake
- 43:01Systems access
- 46:41Banks
- 48:06Tests
- 50:07Information Sharing
- 50:39Developers need to see into production
- 51:07Monitoring tools
- 51:49Displays
- 52:50Meetings
- 53:51Metrics
- 54:50Capture everything
- 55:38Make proper decisions
- 56:04Share metrics data
- 56:31Change agents
related talks
34mDevops or: How I Learned to Stop Worrying and Love the Ops Team
25mDevOps: Why do we Keep Messing Up?
05mDevOps Is Not a Department
07mTen Ways to Doom Your DevOps
44mDevops - What does devops culture mean for engineers?
01h 02mWe have DevOps. No one was fired. But what to do?!
28mDevOPs Lessons for Frontend Developers
05mDevOps and incident management: A recipe for success