The Three Pillars of Observability

Logs are discrete event records. Structured logging outputs events as machine-readable JSON payloads containing context keys (e.g. `userId`, `requestId`), making them easy to index and search.

Metrics represent numerical values aggregated over time (e.g., CPU utilization, HTTP error rates). They are highly storage-efficient, enabling long-term trend analysis and performance alerting.

In microservice architectures, requests hop across multiple servers. Tracing assigns a unique trace ID that propagates across network API calls, mapping out a visual timeline (span graph) of the full request lifecycle to locate backend bottlenecks.