Chaos Engineering in 2021
summary
Advocates argue chaos engineering becomes useful when it’s boring and routine, not theatrical: run a 4-step experiment (ensure observability, measure steady state, form a hypothesis, run repeatable automations), target failures like increased network latency, CPU exhaustion, expired certs or misconfigured retries and circuit breakers, and test on Kubernetes and containers or single-process legacy systems. Adoption barriers are concrete: 50% cite generating buy-in, 27% cite inadequate training, while dozens of open-source and commercial tools now exist.
AI-written summary · about
chapters (24)
- 0:00Intro
- 0:27Talk
- 1:11Today's talk
- 1:58Reliability
- 2:23Remarkably Reliable
- 2:58Resilience
- 3:322020 - test of resiliency
- 4:11Trends
- 8:13Code/Connectivity/Infrastructure/People
- 12:28What is Chaos Engineering? - Definition
- 13:00Testing
- 13:55Chaos Experiment - ideas
- 14:484 steps to Chaos Experiment
- 16:50Chaos Experiment - Outcomes
- 17:25Innovation Adoption Livecycle
- 18:52(Spoiler alert) Boring is good!
- 20:33Tools
- 21:30What's blocking you when doing Chaos?
- 21:55Myths
- 29:00Getting Chaos on the roadmap
- 30:25Shark vs Burger
- 32:42Book + discount code
- 35:10Let's connect!
- 35:25Thank you!
related talks
27mChaos Engineering: The Path to Reliability
23mMaking Chaos Engineering Boring: Debunking Myths Hampering Adoption
18mChaos Engineering: Break It On Purpose
23mPrerequisites for Chaos Engineering
29mThe Psychology of Chaos Engineering
48mChaos Engineering: The Path to Reliability
26mPsychology of Chaos Engineering
21mThe Psychology of Chaos Engineering