Tracking

Monitoring carrier discovery

What discovery does, what a healthy day looks like, and what to check when a reference will not resolve.

What it is

When someone adds a reference without telling us the carrier, discovery works through ranked candidates until one of them returns a tracking document. It is bounded — a fixed number of attempts, a retry schedule, and a daily budget — because each attempt costs an upstream call.

When a customer says "it will not track"

  1. Ask the copilot to diagnose the reference. That shows every attempt across all accounts, the failure reasons recorded, and the discovery job state in one answer.
  2. Check the job status. Still pending means it has not run out of road yet. Exhausted means every candidate was tried and none worked.
  3. Check whether the reference is even valid. A container number with a bad check digit, or a booking reference typed in as a bill of lading, fails for a reason no amount of retrying fixes.

Reading the fleet view

/admin/carrier-discovery is read-only for ops and that is deliberate: monitoring is the ops job, changing the ranking or the budget is not. A success rate that drops sharply usually means one carrier's feed is down rather than anything about us — check whether the failures share a SCAC before escalating.

Was this helpful?