Beyond the Monkeys Chaos Engineering on the Cloud
summary
Compares centralized vs decentralized chaos with concrete tradeoffs, shows an automated pipeline using Chaos Toolkit, AWS Lambda and ephemeral infra spun by IaC with short‑lived tokens and results stored in a test evidence repo. A stopped ECS task in non‑prod reveals unexpected client errors that self‑heal in ~3 minutes. A game day runs across 3 regions with 13 participants, 14 scenarios and 15 findings that yield architecture fixes, observability alerts and a 50% faster autoscaling scale‑up after tuning.
AI-written summary · about
chapters (14)
- 0:00Introduction
- 1:07Chaos Engineering
- 1:34Chaos: Centralized or Decentralized?
- 2:41Decentralized Chaos
- 7:08Chaos-Ready Environment
- 10:03Chaos-Ready Teams
- 12:22Start Simple! (In non-prod )
- 13:01Simple Experiment Valuable Findings
- 14:32Chaos Test Execution: Design Considerations
- 17:29Chaos Test Execution: Architecture
- 19:43How are faults injected? Test Execution Environment
- 22:16Chaos Experiment: Anatomy
- 24:12Game Day & Findings
- 29:18Goldman Sachs Engineering
related talks
57mChaos Engineering Primer
52mGetting started with Chaos Engineering
45mDeveloping a Chaos Architecture Mindset
39mResilience & Chaos Engineering in the Cloud
37mCreating chaos in containers
26mContinuous Chaos in DevOps
49mChaos Engineering: Withstanding Turbulent Conditions in Production
46mEmbracing Collaborative Chaos: Chaos Days on Large Platforms