I’ve Made a Huge Mistake: Implementing Agile on Infrastructure Teams
summary
Attempts to run Scrum on an infrastructure team fail because estimation and user stories break: tickets like import 128 databases into terraform state read as ops tasks, planning poker spits out 1 vs 13 estimate disagreements, and the crucial insight is context gathering is the work. Kubernetes, Terraform, Postfix and cloud quota issues produce constant ad hoc interrupts, sprints overload then idle, demos are worthless, so the team moves to Kanban, adds a stale column and a rota for urgent requests, and measures lead time to improve.
AI-written summary · about
chapters (17)
- 0:00Introduction
- 2:28The Struggle with Planning
- 3:11Work Planning Philosophy
- 4:31Initial Scrum Success
- 6:22Understanding Scrum
- 9:37Scrum in Product Teams
- 13:10DevOps Challenges
- 16:23Scrum on DevOps Teams
- 17:32Inappropriate User Stories
- 19:08Estimation Difficulties
- 20:23Context Gathering is Work
- 21:49Scrum Failure on DevOps
- 23:04Inevitable Unplanned Work
- 26:28Scrum Anti-patterns
- 28:44A Better System: Kanban
- 33:14Managing Unplanned Work
- 35:06Q&A and Conclusion
related talks
29mHow to ruin your product by using Agile as an excuse not to plan
27mApplication-Driven Infrastructure
05mFrAgile - lessons from a failed agile transformation
34mHow to screw Agile and digital transformation
26mNot-So-Greenfeld Agile Implementations
32mTales from the DevOps Transformation Trenches
43mEver Since We Started Doing SCRUM Everything's Great. Not!
48mAgility in Large Transformation Programs – An oxymoron?