Tales from the VOID: The Scary Truth about Incident Metrics
summary
An open database called the VOID collects nearly 2,000 reports for 660 organizations (status pages, blogs, postmortems) and finds duration data are strongly right‑skewed so MTTR is unreliable as a central metric; Monte‑Carlo simulations show small real improvements often vanish in noise. Root Cause Analysis commonly narrows blame and silences people. Alternatives include measuring SLOs and the cost of coordination (people/tools involved), hiring incident analysts, and study near misses to surface what goes right.
AI-written summary · about
chapters (19)
- 0:00Intro
- 0:54Why are we here
- 2:13What are we solving
- 4:27LFI
- 5:04What is the Void
- 6:18Metadata
- 8:50Report
- 9:03Duration Data
- 10:13Distribution of Duration Data
- 11:19The Problem with MTTR
- 12:12Google Incident Metrics
- 18:19Sociotechnical Systems
- 21:22Close Calls
- 22:25Nearmisses
- 23:50Root Cause
- 24:53Path to Blame
- 25:14Matrix Moment
- 25:32New Approach
- 25:51Incident Analysts
related talks
33mThe Problem with MTTR: Learning from Incident Reports
40mFar from the Shallows: The Value of Deeper Incident Analysis
37mIncident Management Metrics That Matter
38mResilience in Modern Software Systems
41mAre We Getting Better Yet? Progress Toward Safer Operations
42mThe Fault in Our Metrics: Rethinking How We Measure Detection & Response
24mThe Metrics Question: Evolving Past MTTX
23mYou Get What You Measure—Why Metrics Are Important