observability

Enable debugPerformance in createRouter or call ctx.debugPerformance() in middleware for selected requests. Rango prints a shared-axis waterfall and adds a Server-Timing header showing handler, render, loader, cache, and middleware spans so the slow phase is visible.

Pass createOTelTracing(tracer) to the tracing slot for phase spans and createOTelSink(tracer) to the telemetry slot for discrete events like handler errors and cache decisions. Both come from @rangojs/router and use a tracer from @opentelemetry/api.

Yes. Use createCloudflareTracing from @rangojs/router/cloudflare to emit phases as native Cloudflare custom spans with no OpenTelemetry dependency. Note that Server-Timing durations read near zero on deployed Workers because timers are frozen during request execution.

On production Cloudflare Workers, timers are frozen during request execution as a Spectre mitigation, so durations only advance across awaited I/O. Use the waterfall locally and rely on client-side timing plus structured telemetry for deployed measurement.

Enable debugPerformance together with a telemetry sink, then inspect revalidation.decision events to see which segment re-ran or skipped, cache.decision events for hit/miss/stale behavior, and loader spans to confirm loaders overlap rendering.

No. debugPerformance is off by default and telemetry emits nothing unless a sink is configured. Off-platform tracing calls are transparent pass-throughs, so requests behave as if tracing were off.