Skip to main content

🧩 Observability

✅ Overview​

Design philosophy to make it possible to grasp what is happening inside the system "from outside".

✅ Problems Addressed​

  • Cannot understand cause at failure (Black box).
  • Tracing in distributed system is difficult.
  • Cannot see improvement points due to lack of data.
  • Increase in complexity where traditional "Monitoring" is insufficient.

✅ Basic Philosophy & Rules​

  • Metrics
  • Logs
  • Traces
  • Integration as "Three Pillars".
  • Data output with context (Trace ID).
  • Incorporate observability into design at code/infrastructure layer.

✅ Suitable Applications​

❌ Unsuitable Cases​

  • Single node small-scale app (Over-engineering).
  • Environment where log management / observability tool operation is impossible.

✅ History​

  • Evolved from Monitoring to Observability.
  • Standardization progressed from Dapper → OpenTracing → OpenTelemetry.

✅ Related Styles​

  • SRE: Observability is foundation of SRE practices.
  • DevOps: Prerequisite for safe acceleration of delivery.
  • Service Mesh: Automation of trace/metrics collection.

✅ Summary​

Observability is an essential approach to keep complex systems in understandable state,
and is indispensable for failure response, improvement, and rapid development.