rust-observability

Structured logging in Rust uses tracing_subscriber to emit JSON logs with context. Instrument functions with #[instrument] macros to create automatic spans, then configure a tracing subscriber to capture and export logs with full trace context for debugging distributed requests.

Yes. OpenTelemetry collects traces and metrics from your Rust application, while Prometheus exporters expose those metrics for scraping. Configure both together to unify tracing, metrics, and structured logs in one observability pipeline.

Correlation context propagation embeds trace IDs in request headers so spans from different services link together. Implement this by extracting parent trace context on ingress and injecting it into outbound requests, enabling end-to-end request tracking.

Apply the #[instrument] macro from the tracing crate to functions you want to trace. This automatically creates spans, captures function arguments as fields, and records their execution time, giving visibility into hot paths and error sources.

Health checks complement observability by exposing service status on a dedicated endpoint. Implement them alongside tracing and metrics to enable orchestrators to detect failures faster and route traffic away from degraded instances.

Structured logs alone don't show request flow across services or execution context. Distributed tracing with OpenTelemetry adds causal relationships, timing data, and automatic span correlation, turning isolated logs into a coherent request timeline.