Don't restrict your living documentation to a living documentation
summary
Living documentation pairs Gherkin scenarios with business metrics, APM and Amplitude so deployed features are measured for value and harm. A previously added “push results back” capability suddenly clogs Sidekiq and makes requests take up to 16 seconds and background queues fill for minutes, revealing a global UX regression. A “trash restore” feature shows 70–90% of restores happen within 1–3 days, guiding whether to keep, promote or remove the feature.
AI-written summary · about
chapters (8)
- 0:00Introduction to BDD
- 0:48Continuous feedback loops
- 2:11Three amigos and automation
- 3:33Living documentation at scale
- 4:42Learning from past failures
- 7:00Predictive performance metrics
- 9:55Analyzing feature impact
- 12:40Data-driven product health
related talks
05mWhy Living Documentation is Cool
10mBeyond “try harder”: Effective strategies to tackle bugs
50m"Living Documentation": When Our Product Definition, Our Automated Tests...
32mOn the Care and Feeding of Feedback Cycles
23mSquash Production Defects Quickly - The Power of Structured Logging in Rails
51mQ&A Session
34mTesting legacy code when you dislike tests (and legacy code)
05mFunctional Documentation Testing