How To Measure MTTx Values in SRE
summary
Explains how to measure MTTF/MTTD/MTTR as latencies (longer MTBF better), why averages hide worst cases and why teams must use percentiles and HDR histogram for accurate latency distribution; shows tracking by revenue impact with tiered customer SLIs and by incident severity; prescribes concrete governance: quarterly governance reviews, toil-backlog triage, critical-dependency plans (Zookeeper and Starbucks Mobile Order examples), SLO-led decisions and cost‑per‑incident accounting.
AI-written summary · about
chapters (16)
- 0:00Introduction
- 1:26SRE motivation journey
- 3:53MTTX metrics defined
- 8:06Risks of MTTX misuse
- 10:55Incident tracking dashboards
- 13:00Latency with percentiles
- 14:47Customer tiering
- 16:57MTBF analysis over time
- 20:08SRE metrics governance
- 22:16Dependency reviews and mitigation
- 24:46Quarterly SRE activities
- 27:12Services vs customer experience
- 30:24SRE as product discipline
- 33:12Product owner collaboration
- 37:02Summary
- 37:44Q&A
related talks
33mThe Problem with MTTR: Learning from Incident Reports
24mThe Metrics Question: Evolving Past MTTX
37mIncident Management Metrics That Matter
38mResilience in Modern Software Systems
38mImplementing SRE practices: SLI/SLO deep dive
29mNot All Minutes Are Equal: The Secret behind SLO Adoption Failure
30mSeeing RED
37mKnow Thy Enemy: How to Prioritize and Communicate Risks