Cloudflare Workers OTel + utels boundary

Two wrappers that compose around a Worker's {fetch, scheduled} handler. Both are no-op pass-throughs unless their env vars are present, so you can run with neither, just one, or both.

withUtelsErrorTracking(withTelemetry(coreHandler))

• withTelemetry — OTLP traces / metrics / logs push when any OTEL_EXPORTER_OTLP_* endpoint is configured. Also wraps every D1 binding with a Proxy that logs event: "d1.slow_query" to wrangler tail, even without OTLP, so the slow-query story works on a fresh deploy.
• withUtelsErrorTracking — pushes one exception event per 5xx response or thrown exception to a utels.dev project. The endpoint, project ID, and ingest token are env-configured.

When to invoke

Use when you're:

• Standing up observability on a new Worker, want OTLP-compatible traces and metrics to any backend (Honeycomb, Grafana Cloud, Tempo, Jaeger collector, …).
• Adding server-side error tracking via utels without changing handler code.
• Investigating a slow query: drop the threshold env var and watch wrangler tail.

What's in here