Health and inventory

Kowan's own health page (databases, disk, host latency, background loops) and the inventory of what runs where, as CSV or JSON.

Health

Settings → Health (admins) shows how Kowan itself is doing, refreshed every 10 seconds:

• Server: version, running since, the size of its databases and the free space in its data folder.
• Hosts: online or not, the round trip of a ping (through the agent's tunnel for a remote host), last seen, and whether crash snapshots follow it.
• Background loops: the monitor, auto-update, uptime checks and log retention, when each last ran and whether it is late (it missed three of its runs).
• Work in flight: runbooks running, images being pulled, open WebSockets.

At the top, Needs a look lists what is wrong: a late loop, an offline host, a disk almost full. The same data is at GET /api/v1/system/health.

Inventory

Workloads → More → Export inventory downloads a CSV (or JSON) with one row per container on every host you can see: host and its groups, workload and its kind, service, container, image and image id, state, published ports, creation time and the deployment that owns it.

From the API: GET /api/v1/inventory?format=csv (or json), with &group=production for one host group. The JSON lists the hosts that could not be read as well.