Week 11 · Observability & break-glass · about 3 minutes
What you will be able to do
Carry a trace ID across Helix gateway, retrieval, and model call so one request is debuggable.
The idea
Mint a trace_id at the edge and propagate it on every hop and log line.
Without it, incidents become archaeology.
trace_id = request.headers.get("X-Trace-Id") or new_id()
log.info({"trace_id": trace_id, "event": "retrieve"})One request, one id, every service.
Key takeaways
- Create at the edge.
- Propagate everywhere.
- Log it on every event.
Part of the Forward Deployed Engineer roadmap. New to the role? Start with what a Forward Deployed Engineer does.