Fear of the computer
summary
A recipe for safe change: add instrumentation to get baselines, use feature flags and rollouts (ramp up by percentage of hosts or internal dogfooding) so regressions can be rolled back quickly, and slice metrics so P95/P99 show edge cases. Concrete failures include a recursive ORM dependency causing out‑of‑memory and a missing timer that hid an API slowdown; both are fixed once discovered, proving the techniques actually work.
AI-written summary · about
chapters (9)
- 0:00Start
- 4:05Techniques for confidence
- 4:23Case Study 1 E
- 5:26Instrument the code
- 6:37Why feature flags for Infrastructure changes?
- 10:14Ramp up incrementally
- 14:22Case Study 2
- 17:16But what's going on?
- 20:02Make big infrastructure changes with confidence
related talks
01h 15mTaking Control Of Your Test Environment
40mBullet-proof projects
19mThink Prod Up Front: Shifting Left on Production Excellence
44mThe role of catastrophic failure in software design
47mHow to win in software development with tighter feedback loops
46mFlaky, Fragile, and Forgotten: The Fall of a Test Automation Project
35mQuickly and Safely Refactoring Legacy Code
35mStart me up, safe!