The Road to Reliability
summary
Claims that Chaos Engineering plus GameDays improve tempo (deployment frequency, lead time) and stability (MTTR, change failure rate) by testing real failure modes using Gremlin against the Google Cloud open‑source Bank of Anthos app. Demo runs black hole attacks on single‑pod and two‑replica services, shows one‑pod balance reader yields a visible “--” balance and failed payments while deposits succeed, and a 50% blast radius on transaction history causes a brief invisible failover that works. Also covers autoscaling spikes (JMeter/Gatling), capacity caps, CI/CD gating, and Gremlin certification.
AI-written summary · about
chapters (12)
- 0:00Reliability and chaos intro
- 1:04Why reliability needs effort
- 4:17Measuring tempo and stability
- 7:06Chaos engineering basics
- 9:37Scientific method for chaos tests
- 13:52Starting small with chaos
- 14:14Live chaos demo on Bank of Anthos
- 20:37Mitigating failures with redundancy
- 23:03Game days for reliability
- 25:50How to run effective game day sessions
- 28:41Automating chaos in CI/CD
- 29:01Free chaos certification overview
related talks
26mBreaking Systems for Resilience
48mChaos Engineering: The Path to Reliability
21mSelf-Service Chaos Engineering: Fitting Gremlin into Grubhub's DevOps Culture
20mKeynote: Finding the Joy in Chaos Engineering
45mGameDays: Practice Thoughtful Chaos Engineering
26mFrom experimentation to continuous verification
45mDeveloping a Chaos Architecture Mindset
01h 02mPractical Chaos Engineering breaking things on purpose to make them more resilient against failure