How did our DevOps team become another silo?
summary
Argues that DevOps becomes a new silo unless engineering embeds practices rather than people: the organization moves developers into teams and evolves them into site reliability engineers, automates deployments through APIs using Docker and Kubernetes, and enforces schema migrations via tools so databases stay backward compatible for blue‑green and canary releases. Uses DORA metrics (deploy frequency, lead time, MTTR, change failure rate) to show high performers deploy hours apart with lower failure rates. Stresses Prometheus-style observability, human ops and cultural change.
AI-written summary · about
chapters (30)
- 0:00Intro
- 0:26Agenda
- 1:38Who am I
- 4:32My 5yearold broke his arm
- 6:12What are silos
- 7:31Martin Fowler
- 8:15How do we push past this
- 9:10KPIs
- 10:50Medium performers
- 11:46Lead time
- 12:14Change failure rate
- 13:17Elite performance
- 13:55Manual work
- 14:19The J curve
- 17:14Test automation
- 18:25Database automation
- 19:45Database migration
- 22:04Bluegreen deployment
- 25:30Alerts and monitoring
- 26:48Autodiscovery
- 28:28Code Prometheus
- 29:00Machine Learning and AI
- 29:43The Westrom model
- 30:43Information actively sought
- 31:35Responsibilities shared
- 32:05Crossfunctional collaboration
- 33:14Human Ops
- 34:24Empathy
- 36:25Code of Conduct
- 36:38Monitoring Platform
related talks
05mDevOps Is Not a Department
41mDevOps for Java Developers (or maybe against them?!)
30mWhat's It Like When DevOps Is Just Another Day at the Office?
47mDevOps is for Product Engineers, too.
24mHow DevOps changed the way we operate software
36mHow DevOps is evolving practice
19m10 years of DevOps - Where Are We Today?
41mThe Introduction of DevOps in Large-Scale Organizations