Observability at the Edge: Instrumenting WebAssembly with OpenTelemetry
summary
Adds end-to-end OpenTelemetry to wasmCloud running on Rust using the otel Rust SDK plus the tracing crate, propagating spans via NATS headers and exporting traces, logs and metrics to standard collectors; traces end at component calls (no function‑level spans yet) so call latency is visible — one example is 81.8 µs — and sampling is required; metrics include machine and lattice IDs by attaching resource labels to every metric; demonstrated working on a wildly distributed wasmCloud deployment.
AI-written summary · about
chapters (12)
- 0:00Introduction
- 1:06What is WebAssembly and why it matters
- 4:57WasmCloud platform
- 6:20OpenTelemetry for observability
- 8:47Rust SDK integration
- 11:16Decision on distributed tracing
- 13:23Tracing in WasmCloud
- 16:30Tracing limitations
- 19:22Adding logs
- 20:03Metrics in Rust
- 22:49Future work
- 23:54Resources
related talks
27mObserving the Future: Embracing OTEL in WebAssembly
35mExploring the Landscape for Open Telemetry for Wasm
32mThe Amazing Trace: Enhancing Distributed WebAssembly With OTEL- Brooks Townsend
28mUnlocking Observability in WebAssembly with OpenTelemetry
35mPiercing the Veil of WebAssembly in Production
23mObservability Without Borders: The OpenTelemetry Collector in a WebA...
55mCloud-Native Observability with OpenTelemetry
27mEnd-to-end visibility in the cloud w/OpenTelemetry