The service visit lifecycle

A service visit is one stop

A service visit is a single machine's stop at a single property during a storm — the atomic unit of work the swarm claims and the operator completes. Every property in an active pass produces a service visit, and following its status is how you know exactly where a property stands.

The stages

1. Pending — the property is live work, waiting for a machine to claim it.
2. Claimed — a run has taken it; it's that operator's next stop. The claim is database-atomic, so two machines can never grab the same lot.
3. On the property (arrived) — the machine's GPS has crossed the property's geofence; the crew is working it.
4. Completed (or skipped) — the work is confirmed done, or the stop was passed with a reason. Completion can fire the customer's service-complete alert.

Claims aren't forever

A claim carries a time-to-live scaled to how far the work is. If a machine goes silent past its window, a background job releases the claim and the property returns to pending for the swarm to reassign — the self-healing that keeps a stalled run from sitting on work no one's doing.

Where you watch it

• The dispatcher dashboard shows every visit's status and flags trouble — a stranded stop (pending work with no live machine to take it) or a stuck claim.
• The customer's live view turns the same lifecycle into a moving ETA and before/during/after updates.

Understanding these four stages is understanding the whole storm: everything the fleet does is service visits moving through them.