SLOs as an Organisational 'Check Engine' Light
Talkat Observability Day 2025 · on Tuesday, 2025-04-01· at 13:30· Level 1 | Hall Entrance S10 | Room B·
summary
A team with an 18,000+ CI test suite for a monolith builds glue services and uses ReportPortal to cluster failures, runs a Monte Carlo analysis showing only ~15 imperfect tests allowed for 80% success, and defines an SLO of 99.97% test pass on main over a 28-day rolling window. The alert acts as a check‑engine light that sometimes forces action and sometimes permits deliberate non‑intervention; overall pass rate sees only intermittent improvements and organisational priorities on profitability limit enforcement.
AI-written summary · about
chapters (11)
- 0:00Introduction
- 1:06Problem context in 2020
- 3:11Testing cognitive bias
- 4:15Failed test fixes
- 6:19Tech solutions limits
- 8:26SLOs to drive change
- 9:51Misaligned goals
- 11:14Rasmus model priorities
- 12:38SLOs as handshake
- 14:23Q&A team SLOs
- 15:46Q&A fire alarm
related talks
32m9 Things You Should Do When Starting to Use SLOs
47mCultivating Production Excellence
37mReal World SLOs and SLIs: A Deep Dive
33mHow and why we lowered our SLO
23mSLO's for Openstack - the API for Engineering Teams
23mGoing from 30 to 30 Million SLOs
29mMeasuring Reliability: What Got Us Here Won't Get Us There
29mReducing Trauma in Production