// Betrieb
Viewing service health & logs
The Services page (nav: Services) is where you see the control plane's own health and every agent reporting into it. It has two tabs.
Services tab
Each hub service (auth-svc, inventory-svc, ingestion-svc, alerting-svc,
gateway-svc) is a compact row — a status dot, its name, and an inline
endpoint · up <uptime> · <memory> summary (with a red N 5xx chip if it's
returning server errors). Click a row to expand it for the full detail —
endpoint (host:port), version, uptime, memory, CPU, and requests broken down by
HTTP status — and a Logs button.
These figures come from each service's /metrics, scraped into VictoriaMetrics
under the meta location (see
Observability & self-monitoring).
Service logs
Logs opens a live tail of that service's log, pulled from Loki through the gateway's authenticated proxy. Pick a time window (15m–24h) and Refresh; lines are coloured by level.
This is hub-service logs only. A spoke-agent's only channel to the hub is
metrics/inventory over NATS — there is no log subject — so agent logs are not
shipped to the hub. Read those on the remote host with docker logs /
kubectl logs.
Agents tab
Every location an agent has reported for, as a compact row (status · resource count · messages · last seen). Expand one for:
- a source breakdown — how many resources each collector (Proxmox, iDRAC, Kubernetes, self) is reporting;
- the live NATS connection from the broker's
/connz: RTT, connection uptime, client, throughput, and subscription count. Agents tunnel NATS over wss through the single public edge, so the hub peer IP shown is the edge, not the agent's own address; - Open location / Onboard / Offboard actions (see Onboarding a new location and Offboarding an agent).
The same agent card appears on a location's own Agents tab at
/locations/<id>, scoped to that one agent.