Cleaning the Sewage out of your DevOps Pipeline
summary
Argues that failing DevOps usually means a broken pipeline not a broken idea and shows concrete fixes: add a feedback loop with Application Insights or Raygun and structured logging (Serilog/Seq); run CI on every commit, prefer trunk-based development or short feature branches and use feature flags; build once and deploy many with Octopus Deploy (avoid rebuilding for prod) and use migration tools like ReadyRoll/EF; apply red‑green‑refactor tests, automate deployments, remove human gates, and expect more failures early (DORA studied ~27,000 orgs).
AI-written summary · about
chapters (8)
- 0:00Start
- 13:00Branching strategies
- 26:13Build once, deploy many
- 27:27Continuous Integration Build server
- 34:25Computers are better at scripts
- 43:04Automate deployments
- 48:48Deploy everything every time
- 52:21One problem at a time
related talks
28mThe DevOps Pipeline
11mDevOps Pitfalls, Beyond The Technology
55mPatterns & Anti Patterns for Scaling & Enabling DevOps Adoption
29mMeasuring Devops Success with Pipeline Analytics
25mDevOps: Why do we Keep Messing Up?
01h 02mThe Enterprise DevOps Challenge
28mDevOps Stories; Think outside of your Pipelines
24mBringing Product Thinking into DevOps