Parity in Observability Across Different Platforms
summary
Riot Games operates 300 Kubernetes clusters with 20,000 deployments and 150,000 pods and builds an automated pipeline that parses observability definitions, applies lookup tables, performs ML-assisted query conversion, validates results and publishes via GitOps to publishers. Early rollout saves over 100 engineering hours, improves publisher observability coverage from under 10% to about 50%, and reduces mean time to detect by 300% for covered services. The team standardizes on OpenTelemetry and governs vendor features.
AI-written summary · about
chapters (14)
- 0:00Observability at Riot Games
- 2:47Observability fragmentation
- 4:31Partial coverage impact
- 6:38Data parity standards
- 9:46OpenTelemetry for parity
- 11:36Vendor-specific governance
- 12:39Why operational parity matters
- 13:21Manual parity limitations
- 14:45Observability in service definitions
- 16:08Infrastructure for parity
- 18:57Automated query translation
- 21:00ML query conversion
- 22:04Results and benefits
- 23:48Publisher experience and plans
related talks
26mRiot Games: Evolution of Observability at the Gaming Company
11mKubernetes Observability
45mDecentralizing SaaS Monitoring Using Open-Source Tools and Standards
13mNext-Gen Observability: reduce MTTR
27mRethinking Observability for Off-the-Shelf Tools
33mBeyond Monitoring: The Rise of Observability Platform
15mImportance of Observability in Platform Engineering
40mSurvivability of Observability: Taking your DIY system to success