Resilience Is Not Enough
summary
Argues that redundancy alone fails and systems must be built to recover quickly and predictably, not just tolerate faults. Recommends concrete practices: fault injection and game-day testing, formal methods and cellular architectures to reduce risk, and a documented correction of errors process with time‑bound action items. Operational choices include making every change reversible with durable audit logs, supporting rapid zero-change deployments or controlled slow‑start ramps, and designing cold‑start and throttling behaviors so restarts succeed under load.
AI-written summary · about
chapters (23)
- 0:00Introduction
- 1:30History of Engineering
- 2:20Engineering Practices
- 4:50Engine Failure
- 5:35Engine Explosion
- 6:30Recovery
- 8:10Finding the root cause
- 8:45Building systems easy to recover
- 9:28Turn off and on again
- 10:06Tetris analogy
- 10:43Complex systems
- 11:25Resilient systems
- 12:12Intentional engineering choices
- 12:36Use the same system
- 13:22One of the most memorable failures
- 14:48Warning signs
- 15:50Waiting for recovery
- 16:32Light jogging
- 17:30Resource exhaustion
- 18:45Change in complex failures
- 19:26Change in complex systems
- 21:08Key takeaways
- 21:39Conclusion
related talks
51mWithout Resilience, Nothing Else Matters
11mResilience engineering mythbusting
05mWhat is resilience engineering
43mResilience Engineering Mythbusting
16mResilient by Design: Architectures for Crisis Response & Recovery
10mDesigning for Failure using AWS: Resilient Systems
05mBuilding resilience in our organizations...
36mDesigning for failure - architecting fault-tolerant system