SLO's for Openstack - the API for Engineering Teams
summary
OpenStack and Kubernetes run across five data centers with over 200,000 cores and the team shows a live demo that injects load with go-gena and measures CPU with noble9 (Prometheus/Thanos hooks supported). SLOs are defined as proportions (example 99.9 → error budget), enacted at the 99th percentile for CPU usage, and trigger preemptive actions (add capacity, rollback) when thresholds begin to trend up, preventing customer-facing outages and keeping engineering cost versus reliability explicit.
AI-written summary · about
chapters (15)
- 0:00Intro
- 0:25Reliability
- 1:09Error Budget
- 2:31SLO Definition
- 3:23Happiness Test
- 4:14Errors Budget
- 5:08Service Dependencies
- 6:11Openstack SLOs
- 9:28Demonstration
- 12:04Evolution of SLO culture
- 15:37Business technical feedback loop
- 15:56SLOs for the business
- 18:03How to calculate SLOs
- 21:24Feedback loop
- 22:05Summary
related talks
37mUsing SLOs for Continuous Performance Optimizations of Your K8s Workloads
31mServerless Observability: a case study of SLOs
31mSLX: An Extended SLO Framework to Expedite Incident Recovery
33mHow and why we lowered our SLO
18mSLOs as an Organisational 'Check Engine' Light
01h 01mNon Functional Requirements For The Win!
32m9 Things You Should Do When Starting to Use SLOs
30mFrom Monitoring to Observability: Left Shift your SLOs with Chaos