Chaos Engineering: The art of breaking stuff
summary
Argues that resilience comes from planned experiments not monitoring alone and demonstrates a workflow: establish steady state, form a hypothesis, design and execute a reversible experiment, learn and fix. Recommends routing a small percent of users, expanding the blast radius, running game days and continuous chaos. Names tools: Gremlin, Chaos Toolkit, Chaos Monkey/Simian Army, kube-monkey. Cites Netflix running ~1,500 automated experiments/day and warns hard rollbacks must be fast and predictable.
AI-written summary · about
chapters (11)
- 0:00Introduction and speaker background
- 1:53Why chaos engineering matters
- 4:26Complexity of modern cloud systems
- 7:01Measuring reliability beyond uptime
- 12:30Failure is inevitable and how to handle it
- 16:04Chaos engineering as experimentation
- 21:48How to start with chaos engineering
- 29:19Designing and executing experiments
- 34:01Learning and improving from experiments
- 37:15Architecture patterns for resilience
- 47:08Key takeaways and closing
related talks
52mChaos engineering: Breaking things on purpose in production
43mBuilding a culture from Chaos
46mDesigning a System for Chaos
33mChaos Engineering, how to increase resilience by injecting chaos
57mChaos Engineering Primer
38mChaos Engineering: A Step Towards Resilience
18mFostering Chaos Engineering Culture
49mChaos Engineering: Withstanding Turbulent Conditions in Production