System Information and Instances

The System Information page is for Root administrators. It provides instance status, resource usage, and system-task information for multi-node deployments. Open System Information from the profile menu or visit /system-info directly.

Public Health Checks

The platform provides three unauthenticated GET endpoints for Kubernetes, load balancers, and external monitoring:

If the database probe fails or times out, /health/ready returns HTTP 503:

{
  "success": false,
  "status": "unhealthy",
  "message": "database connection failed"
}

Clients receive only the generic message; the specific database error remains in server logs. Orchestrators should use /health/live to decide whether to restart a container and /health/ready to decide whether to send it traffic. Do not parse an administrator-only System Information API as a readiness substitute.

livenessProbe:
  httpGet:
    path: /health/live
    port: http
readinessProbe:
  httpGet:
    path: /health/ready
    port: http

Instance Status

The instance list shows the node name, role, version, start time, last heartbeat, and resource status. Instances report their status every 30 seconds by default. A record whose last heartbeat is more than 90 seconds old is marked as stale.

Configure a stable, unique NODE_NAME for each node so the same instance remains identifiable during rolling updates and asynchronous-task troubleshooting.

Remove Stale Instances

Only Root administrators can remove instance records.

1. In System Information -> Instances, confirm the stale records to remove.
2. When removing one record, the backend rejects instances that are online or still within the stale threshold.
3. Use Delete all stale to remove every stale record in one action. The list refreshes after confirmation.

Bulk removal only affects records whose last_seen_at is earlier than the current time minus the stale threshold. Online instances and records exactly on the threshold are not removed. Refresh and single-instance removal are temporarily disabled while the operation is running to prevent duplicate actions.

Management API

The Root-only endpoint is DELETE /api/system-info/stale-instances. On success, data.deleted_count reports the number of records actually removed. The operation reuses the existing system_instances data and requires no database migration.

Removing an instance record does not stop a process, revoke keys, or change channel configuration. Investigate networking, clock synchronization, and heartbeat reporting before removing a record for a node that might still be running.