Continuous Chaos in DevOps
summary
Capital One runs Cloud Detour, a serverless (AWS Lambda) chaos-engineering automation that registers apps, exposes CLI/API, integrates with CI/CD, enforces approval/audit workflow and runs health checks. It simulates terminations, process kills, EBS/ECS/EBS network faults, S3 policy blocks, RDS failovers and region-wide outages. Pilots on 50+ apps uncover unexpected single-process failures, degraded API/S3 paths and slow failover timing; results feed design fixes and a planned open‑source release in Q1 2018.
AI-written summary · about
chapters (11)
- 0:00Intro
- 0:27Where we are today
- 3:12Everything Frames
- 4:13Navigating Complexity
- 5:03Antifragility
- 6:38Embrace failures
- 9:06Chaos Engineering
- 11:38Cloud Eater
- 18:04Features
- 22:29AWS Cloud
- 24:19Final Thoughts
related talks
57mChaos Engineering Primer
29mImprove Resilience with automated chaos engineering
33mBeyond the Monkeys Chaos Engineering on the Cloud
27mPutting Chaos Into Continuous Delivery to Increase App...
39mResilience & Chaos Engineering in the Cloud
46mEmbracing Collaborative Chaos: Chaos Days on Large Platforms
35mImprove resilience with automated chaos engineering
22mAll Clouds Need a Sprinkle of Chaos
