Machines

Machines are the compute nodes that back your workloads — the individual servers in your Kubernetes clusters, across regions. There is no separate machine pool to manage: machines are projected directly from your cluster node pools.

Machines Are Cluster Nodes

Every machine belongs to a cluster — either shared Hanzo Cloud or your own dedicated DigitalOcean Kubernetes (DOKS). A cluster declares one or more node pools (a machine size and a count); the operator reconciles those pools into running nodes, and the Machines page projects each node from its pool.

GPU node pools additionally surface as GPUs; everything else is general-purpose CPU and memory capacity.

Regions

Clusters — and therefore machines — are placed in a region close to your users:

Inventory API

List your clusters and their node pools — the source of truth for machine capacity — through the PaaS control plane. Replace {org} with your organization slug:

curl https://api.hanzo.ai/v1/paas/org/{org}/cluster \
  -H "Authorization: Bearer hk-..."

{
  "clusters": [
    {
      "id": "prod-sfo3",
      "region": "sfo3",
      "nodePools": [
        { "size": "s-4vcpu-8gb", "count": 3 },
        { "size": "gpu-h100x1-80gb", "count": 2 }
      ]
    }
  ]
}

To see what is actually running on those machines, list the deployed apps and their health:

curl https://api.hanzo.ai/v1/paas/apps \
  -H "Authorization: Bearer hk-..."

Related

• GPUs — accelerated machines and their telemetry
• Pipelines — deploy workloads onto your machines
• Edge — run inference off-cluster, on-device
• API Keys · API Reference