fleet-manager: host uptime metric + reboot-pending status

The manager lets an admin schedule a reboot (desired.reboot_at) but never showed
whether it had happened. Add host uptime to the health snapshot and derive a
reboot status from it.

- proto: HealthSnapshot.uptime_seconds (15); regenerated pb.go (manager + agent copy).
- agent: probe /proc/uptime -> uptime_seconds (with parser + test).
- manager: Health.UptimeSeconds cached in inventory; snapshotToHealth maps it;
  influx writes + whitelists uptime_seconds as a queryable field.
- UI: fmt.js dur() + rebootStatus() — boot_time = last_seen - uptime (manager clock
  frame, no host/manager sync needed), compared to reboot_at. Monitor host detail
  shows an "uptime" row and a "reboot" row: pending in X / overdue by X / rebooted.

Observability-only — works before the reboot ACTUATOR lands (deferred). Once a host
boots after the scheduled time (BMC/manual now, agent later), the badge flips to
"rebooted" automatically.

serving/fleet-manager/influx.go

@@ -26,6 +26,7 @@ const influxMeasurement = "fleet_health"
 var metricFields = map[string]bool{
 	"load_avg": true, "cpu_pct": true, "temp_c": true, "mem_used_pct": true,
 	"tmpfs_used_pct": true, "chromium_crashes": true, "kiosk_active": true,
+	"uptime_seconds": true,
 }
 
 var serialRe = regexp.MustCompile(`^[0-9a-fA-F]+$`) // cert serials are hex
@@ -80,6 +81,7 @@ func snapshotPoint(serial, hostnqn, starship string, snap *fleetpb.HealthSnapsho
 		"tmpfs_used_pct":   snap.GetTmpfsUsedPct(),
 		"chromium_crashes": int64(snap.GetChromiumCrashes()),
 		"kiosk_active":     boolToInt(snap.GetKioskActive()),
+		"uptime_seconds":   snap.GetUptimeSeconds(),
 		// string status fields kept as fields too (handy for annotating charts)
 		"nvme_root": snap.GetNvmeRoot(),
 		"auth":      snap.GetAuth(),