There is No Root Cause...
summary
Argues that outages rarely have a single cause and proposes replacing RCA with the "can even" framework (four quadrants: simple complicated complex chaotic) plus practices to manage emergence. Cites microservice/AWS architectures, feature flags, and Victor ops incident management platform as context, recommends game days and probing/sensing/responding to move patterns toward simpler quadrants. Includes concrete examples of responses per quadrant and guidance for postincident reviews and roadmap work.
AI-written summary · about
chapters (32)
- 0:00Introduction
- 1:07VictorOps
- 1:33Root Cause
- 2:22Fishbone Diagram
- 3:10Root Cause Analysis
- 3:47Root Cause Criticism 1
- 6:33Root Cause Criticism 5
- 7:46Complex Architecture
- 8:11Change Vectors
- 8:55Change vector
- 9:16The Internet
- 9:45We have all been present
- 10:05Our systems are trees
- 10:21Emergence
- 10:59Systems
- 11:31Root Cause Language
- 12:13New Patterns
- 12:26Are we doomed
- 13:05Can Even
- 13:25What is Can Even
- 14:55Simple Quadrant
- 15:39Complicated Quadrant
- 16:32Complex Quadrant
- 17:12Chaotic Quadrant
- 17:53Disorder Quadrant
- 18:49Moving Patterns
- 19:38Retrograde Action
- 20:08Adoption is Not Complex
- 21:34Side by Side
- 23:52There is no broken
- 24:20Voltaire
- 25:14Simple vs Complex
related talks
05mIgnite: Root Cause Analysis Fact and Fiction
19mPrincipled Identification of "Root Causes" Using Techniques...
37mIncidents are a new normal
15mMaintaining reliable systems
28mRevolutionizing Incident Response
13mNavigating SRE/Incident Management
23mDiving through RCAs a quest for reliability
34mIncidents are new normal