Chaos Engineering: The Fine Art of Breaking Stuff in Production on Purpose
summary
Claims concrete failure modes and tools: simian army, Chaos Toolkit, Gremlin; Amazon loses over 100 million for one hour of downtime and a 100 ms latency causes ~1% sales drop. Process shows route 1% of traffic to control and experiment, abort quickly if cascading failures appear, run game days, measure user‑impact metrics (play clicks, sales), start in staging, increase blast radius, rerun after fixes, and embed continuous chaos (kill instances regularly) to build observable, resilient systems.
AI-written summary · about
chapters (8)
- 0:00Start
- 15:11CHAOS ENGINEERING
- 16:19ENGINEERING EXPERIMENTS
- 19:47NT RESPONSE LEARNING
- 20:58EXPERIMENT PHASES
- 24:09HYPOTHESIS
- 26:40EXECUTE EXPERIMENT
- 34:54LTURE CHANGE
related talks
18mChaos Engineering: Break It On Purpose
01h 02mPractical Chaos Engineering breaking things on purpose to make them more resilient against failure
57mChaos Engineering Primer
46mDesigning a System for Chaos
26mBreaking Systems for Resilience
21mThe Psychology of Chaos Engineering
01h 04mHarnessing Chaos; the hidden ingredient behind building better systems through learning
36mChaos Engineering