How to design Observability before you buy it
summary
Claim: projects buy observability platforms and still fight noisy alerts and high MTTR because they skip upfront design. The proposed SRE-driven framework defines personas and use cases, maps capabilities and domains (logs, metrics, traces, ingestion, analytics), sets SLIs and SLOs tied to customer-facing KPIs, and only then chooses tools or combinations plus needed integrations. Also enforces a step zero: align business and technical decision makers so instrumentations and trade-offs (build vs buy, cost vs telemetry depth) actually deliver reduced incidents.
AI-written summary · about
related talks
29mObservability; a gentle introduction
13mObservability that matters
09mSetting up Intelligent Observability
42mDemystifying SRE and DevOps
33mEffective observability in microservice architectures
20mI Can Sell You Observability, But You Can't Buy It
15mEmbedding Platform SREs: A Hybrid Model for Driving Adoption
36mThe road to world-class observability