Not All Minutes Are Equal: The Secret behind SLO Adoption Failure
summary
Claim: minute-by-minute time-slicing hides real reliability signals so time-slicing fails for many services. Using 30‑ and 60‑day traffic and failure logs the team builds error budget remaining charts that correlate with incidents and changes and reveal slow burns and fast burns and silent recoveries. They prototype software to compute and visualize EBR, adopt a bootstrap policy with an EBR default 50%, and document a loop: define scope, instrument, assess baseline, map to customer and add policy.
AI-written summary · about
chapters (10)
- 0:00Introduction and context
- 1:13Defining SRE objectives
- 3:46The math behind reliability
- 6:46Testing measurement strategies
- 10:54The "money chart" discovery
- 14:01Identifying reliability signals
- 17:27Scaling and bootstrapping
- 20:45Socio-technical challenges
- 22:52The iterative SRE process
- 25:16Summary and final thoughts
related talks
38mImplementing SRE practices: SLI/SLO deep dive
47mSRE Jumpstart
06ma pint sized introduction to SLO
29mMeasuring Reliability: What Got Us Here Won't Get Us There
47mCultivating Production Excellence
33mHow and why we lowered our SLO
30mMetrics for good, not evil
09mHow to answer tricky questions the SRE way