Availability Objectives of SoundCloud’s Microservices
summary
SoundCloud sets a minimum availability for each microservice and audits all ~140 services to find risky ones. Availability uses two thresholds: error rate and 99th-percentile latency; metrics are total, failed and slow transactions (slow = above latency threshold). Traffic is intercepted by an SLR that emits metrics to the primitives monitoring system; a services directory fetches metrics, stores history and flags services below their objective. Result: prioritized fixes, fewer pointless optimizations, and data-driven decisions.
AI-written summary · about
chapters (9)
- 0:00Intro
- 0:42Who is SoundCloud
- 1:29How do we improve availability
- 2:25What is serviceable objectives
- 3:56Availability definition
- 5:25Infrastructure overview
- 8:36Graceful degradation
- 9:57Summary
- 12:02QA
