Network Observability with Envoy
summary
Envoy emits listener, cluster, upstream and downstream metrics plus counters, gauges and histograms so dashboards alert on Prometheus pull with histograms for latency percentiles and on connection and SSL cert failures. Use tags (dogstatsd or Prometheus) because statsd lacks tags and raw metric names become unmanageable. Monitor retries, timeouts and outlier ejection monitored separately from downstream errors, and enable the HTTP/1 bridge decodes gRPC codes so gRPC status and methods appear in metrics and traces sent via a Zipkin/collector pipeline.
AI-written summary · about
chapters (10)
- 0:00Introduction to observability
- 1:56Service mesh architecture
- 5:49Defining observability
- 8:16Envoy metrics configuration
- 13:44Metrics formats and collection
- 18:32Analyzing connection metrics
- 24:58gRPC and cluster monitoring
- 29:01Reliability and security
- 32:15Tracing and logging
- 35:45Q&A and closing
related talks
01h 32mHands-on Workshop: Envoy Metrics Demystified Workshop
33mEnvoy on Fire: A Practical Look at Debugging a Service Mesh
22mObservability with Consul Connect
20mNetwork Monitor: A Tale of ACKnowledging an Observability Gap
24mObservability with HashiCorp Consul Connect (Service Mesh)
36mFine Grained Mesh Metrics for Better Visibility With Native Performance
25mGraph-based ML Anomaly Detection and Insights for Envoy Systems
14mMonitoring Network Traffic in Kubernetes