Relationship Between Logging, Metrics, and Tracing

Relationship Between Logging, Metrics, and Tracing An architecture diagram generated by Archify. Logging · "What happened?" · Three Pillars Logging "What happened?" Metrics · "Is the system healthy?" · Three Pillars Metrics "Is the system healthy?" Tracing · "Why is it slow?" · Three Pillars Tracing "Why is it slow?" Event-based · unstructured data · Optimal for debugging · Data shape Event-based · unstructured data Optimal for debugging Time-series · aggregated values · Optimal for alerting · Data shape Time-series · aggregated values Optimal for alerting Request flow · causality analysis · Optimal for performance analysis · Data shape Request flow · causality analysis Optimal for performance analysis Shared labels Exemplars Trace ID Three Pillars Data shape Legend Observability pillar Data shape

Logging

  • • Event recording, auditing, debugging
  • • High data volume, high storage cost

Metrics

  • • System state monitoring, alerting
  • • Medium data volume, sensitive to cardinality

Tracing

  • • Investigate request flow and latency
  • • Evaluate sampling, loss and operating cost