The Three Elements of Observability

The Three Elements of Observability An architecture diagram generated by Archify. Envoy sidecar · Zero-code auto collection · Architecture component · Zero Instrumentation Envoy sidecar Zero-code auto collection Zero Instrumentation Metrics · Requests · latency · errors · 1. Metrics Metrics Requests · latency · errors Prometheus · or VictoriaMetrics · 1. Metrics Prometheus or VictoriaMetrics Distributed Tracing · Request path · dependencies · 2. Distributed Tracing Distributed Tracing Request path · dependencies Jaeger / Zipkin · Tempo long-term storage · 2. Distributed Tracing Jaeger / Zipkin Tempo long-term storage Logging · HTTP calls · security events · 3. Logging Logging HTTP calls · security events Grafana Loki · via Fluentd · or Elasticsearch · 3. Logging Grafana Loki via Fluentd · or Elasticsearch Kiali · Service topology · Integrated observability Kiali Service topology Grafana dashboard · Metrics · traces · logs · Integrated observability Grafana dashboard Metrics · traces · logs Prometheus Alertmanager · Golden Signals alerts · Integrated observability Prometheus Alertmanager Golden Signals alerts metrics spans access logs scrape visualization alert rules 1. Metrics 2. Distributed Tracing 3. Logging Integrated observability Legend Dashboards Signal Backend Alerting Source (Envoy)

Metrics

  • • Measures request count, response time, error rate, and resource utilization
  • • Used for system health monitoring, SLO/SLI tracking, and capacity planning
  • • Tools: Prometheus, Grafana, VictoriaMetrics

Distributed Tracing

  • • Tracks the complete path of a single request and each service's processing time
  • • Used for bottleneck identification, root cause analysis, and microservices debugging
  • • Tools: Jaeger, Zipkin, Grafana Tempo

Logging

  • • Records all HTTP requests/responses, errors, and security events
  • • Used for detailed debugging, security audits, and compliance
  • • Tools: Grafana Loki, Elasticsearch, Fluentd