Ruin Your Systems By Making Them Perfect
summary
Asserts that chasing perfection breaks systems and people and offers concrete practices: measure user-facing metrics as SLIs/SLOs, set an error budget (example: 10 errors per 30 days — a bad deploy causing 20 errors burns two months of budget and triggers a feature freeze), use chaos engineering fault injection to reveal emergent failures, celebrate postmortems and promote those who own fixes, expect about 20% human error on repetitive tasks, accept medical devices run around 99.99% reliability, and budget each extra “nine” at roughly 10x cost.
AI-written summary · about
chapters (35)
- 0:00Intro
- 1:22Overview
- 4:59Human imperfection
- 6:21Embrace our humanity
- 6:49What is reliability
- 8:34Reliability is most important feature
- 9:32Humans are terrible computers
- 12:50Perfection is counterproductive
- 15:50The magic line
- 17:42The error budget
- 18:47Service level indicators
- 24:44Its not your fault
- 26:29We dont blame humans
- 30:14Reliability speed are not enemies
- 31:41Good luck vs bad luck
- 32:58Chaos engineering
- 33:36Distributed systems
- 34:06Humans make wellintentioned mistakes
- 35:08The secret to happiness
- 36:51QA
- 38:01Judgment
- 38:50Autonomy
- 39:18Requirement Capture
- 39:40Question
- 41:19Are pacemakers perfect
- 42:45Target Zero Defect Delivery
- 43:13How to Change Your Mindset
- 44:38What is Acceptable
- 47:39Expectations
- 48:01New System
- 50:13Contingency Planning
- 50:40My Rule of Thumb
- 51:47Customer Feedback
- 52:22Customers Expect Perfection
- 53:45Conclusion
related talks
29mThe Curious SRE: Cultivating a Mindset for System Resilience
06mLightning Talk: Reliability Engineering for Humans
01h 00mHow to be WRONG
39mConfessions of a Systems Engineer: Learning from My 20+ Years of Failure
22mHow to be an inefficient developer?
28mI am bad at things – And that’s okay
14mHow to Succeed with Your Software Project
18mThe Good, the Bad and the Ugly: The 3 Learnings of an SRE