/dashboard/fleet is a read-only health and capacity view of the CXB Core worker pool.
Data and refresh
The page pollsGET /api/v1/fleet/status every 5 seconds. The response contains aggregate fields (total_capacity, active_calls, available) and a workers array. There are no controls on this page — it is observe-only.
Surfaces
| Surface | What it shows | Source field |
|---|---|---|
| Capacity bar | active_calls / total_capacity calls; bar turns red above 80% usage | total_capacity, active_calls |
| Worker card header | Worker URL + Healthy/Unreachable state | worker.url, worker.status |
| Worker capacity row | total_capacity calls (workers count) | worker.total_capacity, worker.workers |
| Worker available row | available slots; red when ≤ 5 | worker.total_available |
| Worker usage bar | used / capacity fraction | derived from total_capacity − total_available |
status === 'healthy', otherwise Unreachable. When workers is empty the page shows “No workers registered”.
Related docs
Fleet routing
How the CXB API picks a fleet server for outbound calls.
Settings
Where fleet URLs are configured.