Cluster API

Health Check

GET /api/health public

Public health endpoint. Returns cluster health status and agent readiness. Also triggers expired session cleanup.

Response:

{
  "status": "healthy",
  "agents": [
    {"name": "my-agent", "ready": true, "status": "Running"}
  ],
  "timestamp": "2026-03-02T12:00:00.000Z"
}

The status field is:

• "healthy" — All agents are ready
• "degraded" — One or more agents are not ready

Cluster Info

GET /api/cluster admin

Node resources, capacity, and allocation details.

Response:

{
  "nodes": [
    {
      "name": "my-node",
      "status": "Ready",
      "capacity": {
        "cpu": "8",
        "memory": "15Gi",
        "pods": "110"
      },
      "allocatable": {
        "cpu": "8",
        "memory": "14.5Gi",
        "pods": "110"
      }
    }
  ]
}

Recommended Infrastructure

Component	Details
Server	Any VPS or bare-metal server
CPU	4+ cores recommended
RAM	8 GB minimum, 16 GB recommended
Disk	50+ GB
k8s	k3s (recommended) or any conformant cluster
Ingress	Traefik (bundled with k3s)
Storage	local-path-provisioner or any CSI driver

Current resource usage per component:

Component	RAM
k3s control plane + Traefik	~160 MB
cloudflared	~18 MB
manyclaws-controller	~80 MB
Per agent (typical)	~520 MB

Comfortable capacity: 5-8 agents on 16 GB RAM. For 10+ agents, upgrade to 32 GB RAM.