Using Istio to Build Better Service Level Objectives
summary
Uses the Istio mesh plus Istio + Prometheus, Grafana, Jaeger, Kiali to generate mesh-to-mesh metrics and traces from a running bookinfo demo on EKS (Kubernetes 1.27, ALB controller) under Locust load. Shows deriving SLOs from observed P90 latencies: product page P90 ≈ 5–7s so define 99% ≤ 6.5s; set upstream targets tighter (example details 99.5%) because nested traces reveal reviews→ratings chains and a spiky reviews v3 that risks end-to-end SLOs.
AI-written summary · about
chapters (11)
- 0:00Istio and SLOs intro
- 3:09SLOs real-world analogy
- 7:02Defining SLOs with Istio
- 9:13Demo with Bookinfo
- 10:58Mesh metrics in Prometheus
- 12:04Kiali for service insights
- 13:48Jaeger for nested tracing
- 15:33Granular mesh metrics
- 18:02Latency SLOs with P90
- 21:56Open-source SLO tracking
- 22:59Summary
related talks
56mObservability in Action for your Service Mesh
30mFrom Monitoring to Observability: Left Shift your SLOs with Chaos
31mServerless Observability: a case study of SLOs
05mIgnite: Istio+: Enhancing Observability and Tracing using Kiali and Jaeger
25mThe Lean Observability Stack: Quick and Native Telemetry for Service...
24mObservability for Istio Service Mesh
40mScaling SLOs with Kubernetes and Cloud Native Observability
25mAn SLO-Driven Approach to Enhance Kubernetes Cluster Reliability