Smart Chaos : Leveraging Generative AI
summary
Claims that generative AI can automate most chaos engineering stages and demonstrates it with an EV charging architecture fed into AWS Party Rock, producing usable dependency extraction, steady-state definition, failure hypotheses, test cases and autonomous chaos workflows. The demo auto-generates service maps, estimates blast radius, integrates with observability and CI/CD, and iterates based on telemetry. It warns that results depend on data quality and template libraries, and that humans still review scope and bias.
AI-written summary · about
chapters (23)
- 0:00intro
- 0:22preamble
- 0:42agenda
- 1:49modern enterprise are complex and increasingly hard to manage
- 3:25next major outage - when?
- 6:00why traditional testing falls short?
- 8:08why chaos engineering
- 10:10what is genai (generative ai)
- 11:05application of genai
- 11:49chaos engineering - "methodology"
- 15:51enhancing chaos engineering with generative ai: a comprehensive framework
- 19:58generative ai in action: illuminating possibilities via architecture diagrams
- 21:17dependency analysis
- 23:48steady state defining
- 24:46hypothesis creation
- 26:06experiment design
- 27:06blast radius defining
- 27:50generative ai in action: illuminating possibilities via auto generated service maps
- 28:51the rumsfeld matrix - "known knowns, known unknowns, and unknown unknowns"
- 30:20known known
- 30:32known unknown
- 34:01target and state
- 35:46best practice and potential pitfalls
related talks
09mAI and Chaos Engr: Smarter Failure Testing for Resilient Systems
27mChaos Engineering: The Path to Reliability
29mImprove Resilience with automated chaos engineering
21mRevolutionizing software creation: Harnessing the power of GenAI
43mTaming the Unpredictable: Reliability in Chaos
18mChaos Engineering: Break It On Purpose
31mTesting, Evaluating and Debugging Generative AI and Agentic Ap...
01h 04mHarnessing Chaos; the hidden ingredient behind building better systems through learning