Continuous delivery in 10 easy mistakes
summary
A practical playbook of pitfalls and fixes: teams try to automate everything without vision, run three‑hour test suites that block pipelines, and accidentally hit 7,448 releases in one day while lacking monitoring and controls. Real fixes include microservices or component splits, visible deploy and reggae boards, tying commits to JIRA, and using feature flags plus canary 5% rollouts. Measured wins: 60% reduction in bugs (250 to 80/month), a ten‑person security satellite, fewer branches, weekly rehearsal deployments, and tracking lead time, release frequency, restore time and change‑fail rate.
AI-written summary · about
chapters (13)
- 0:00Intro
- 2:22Software development is complex
- 8:45What is continuous?
- 12:35Level 1
- 17:35Test pyramid reinvented
- 20:06Maintainable tests
- 21:24The shipping delay pyramid
- 23:29Release strategy
- 29:05Business IT antipatterns
- 31:44Product vision and business cases
- 33:56The red sprint antipattern
- 34:30No estimates movement
- 38:59Legacy code and broken trust
related talks
40mContinuous Delivery Anti-Patterns
46mContinuous delivery antipatterns
55mContinuously Delivering Continuous Delivery
40mSustainable Continuous Delivery: Patterns for rapid and reliable Releases
46mSix small steps towards continuous delivery
05mRelease management as a catalyst for CI/CD
41mShift Left, Shift Right and Improve the Centre
51mBetter software, faster: principles of Continuous Delivery and DevOps