opentelemetry-rust

Tracing and metrics in Rust are added using the OpenTelemetry SDK to instrument code with spans and collectors. Add opentelemetry, opentelemetry-sdk, and opentelemetry-otlp dependencies, initialize a tracer provider at startup, then wrap operations in spans using the tracer's in_span method to capture timing and context automatically.

Yes, OpenTelemetry integrates with Axum, Actix-web, Rocket, and Tonic-based services through manual instrumentation of HTTP handlers and middleware. The Rust SDK provides the APIs to propagate context across requests and record spans for each operation these frameworks handle.

Production-ready OTLP exporters in Rust require setting resource attributes, semantic conventions, proper context propagation across async tasks, and error recording in spans. Initialize the exporter with opentelemetry-otlp, configure the tracer provider globally, and ensure context flows through tokio tasks for complete distributed tracing.

Context propagation across async tasks uses OpenTelemetry's context API to attach trace and span IDs to tokio task spawns. Manually pass context through task boundaries or use instrumentation patterns that preserve the active span when tasks execute concurrently or in sequence.

Yes, the OpenTelemetry Rust SDK supports metrics collection via counters, histograms, and gauges exported through OTLP. Define meters, create instruments for each metric type, record values during application execution, and export them to a metrics backend for aggregation and dashboarding.

Semantic conventions standardize attribute naming and span naming across services, enabling consistent querying and correlation in observability backends. They ensure resource attributes and span metadata follow OpenTelemetry specifications, making traces and metrics portable and searchable across teams.