Engineering Data Reliably Using SLO Theory
summary
Argues that applying SLO theory to data platforms yields measurable wins: define 99.5% freshness over 30 working days (a 90 minutes error budget), then add observability and automated data‑quality gates. The team halves pipeline run time, automates per‑source quality tests, cuts feature lead time in half, and sees support requests reduced by a factor of three, while lowering compute costs — concrete numbers and the implemented pipeline changes back the claim.
AI-written summary · about
chapters (11)
- 0:00Start
- 1:47PATTERNS AND PAINS
- 3:32LONG DELIVERY CYCLES
- 7:41DEVOPS AND SRE
- 9:33SERVICE LEVEL THEORY
- 10:35INDICATORS
- 11:44AGREEMENTS
- 13:29SLOS ARE PRODUCT DECISIONS
- 14:22SLOS SHOULD MAP TO BUSINESS VALUE
- 18:22WHAT IS THE PRODUCT AND WHO IS THE PRODUCT OWNER?
- 19:01BUILD SLOS THAT MAKE SENSE
related talks
28mLean Product Development Through SLOs
30mMetrics for good, not evil
09mHow to answer tricky questions the SRE way
30mTechniques for SLOs and Error Budgets at Scale
31m9 Things You Should Do When Starting to Use SLOs
38mImplementing SRE practices: SLI/SLO deep dive
29mReducing Trauma in Production
33mHow and why we lowered our SLO