fuse metrics

Print fleet-wide Prometheus metrics from the orchestrator

fuse metrics

No flags beyond the globals. Scrapes the orchestrator’s /metrics endpoint and prints the raw Prometheus exposition format, unauthenticated on the server side, but the CLI still sends the active context’s bearer token if one is set (the orchestrator ignores it for this endpoint).

This is fleet-wide, aggregate data: reconcile cycles, tracked VM counts, HTTP request rates and latencies. There is no per-VM metrics API, and for per-host capacity use fuse host metrics <id> instead, this command doesn’t accept a host or VM argument.

$ fuse metrics
# HELP orchestrator_reconcile_cycles_total Total reconcile cycles completed
# TYPE orchestrator_reconcile_cycles_total counter
orchestrator_reconcile_cycles_total 142
# HELP orchestrator_fleet_tracked_vms VMs currently tracked by the fleet manager
# TYPE orchestrator_fleet_tracked_vms gauge
orchestrator_fleet_tracked_vms 3
...

See Monitoring for the full metric reference, what each one means and how to scrape it with Prometheus directly instead of via this command.