Update a GPU cluster

Update the configuration of an existing GPU cluster.

PUT /compute/clusters/{cluster_id}

Together AI SDK (Python)

Python

from together import Together
client = Together()

cluster = client.beta.clusters.update("cluster_id", cluster_type="KUBERNETES", num_gpus=24)
print(cluster)

200

{
  "cluster_id": "<string>",
  "cluster_type": "KUBERNETES",
  "region": "<string>",
  "gpu_type": "H100_SXM",
  "cluster_name": "<string>",
  "volumes": [
    {
      "volume_id": "<string>",
      "volume_name": "<string>",
      "size_tib": 123,
      "status": "<string>"
    }
  ],
  "status": "WaitingForControlPlaneNodes",
  "control_plane_nodes": [
    {
      "node_id": "<string>",
      "status": "<string>",
      "host_name": "<string>",
      "num_cpu_cores": 123,
      "memory_gib": 123,
      "network": "<string>",
      "phase_transitions": [
        {
          "phase": "NODE_PHASE_PENDING",
          "transition_time": "2023-11-07T05:31:56Z"
        }
      ],
      "public_ipv4": "<string>"
    }
  ],
  "gpu_worker_nodes": [
    {
      "node_id": "<string>",
      "status": "<string>",
      "host_name": "<string>",
      "num_cpu_cores": 123,
      "num_gpus": 123,
      "memory_gib": 123,
      "networks": [
        "<string>"
      ],
      "phase_transitions": [
        {
          "phase": "NODE_PHASE_PENDING",
          "transition_time": "2023-11-07T05:31:56Z"
        }
      ],
      "instance_id": "<string>",
      "latest_remediation": {
        "id": "<string>",
        "cluster_id": "<string>",
        "instance_id": "<string>",
        "mode": "REMEDIATION_MODE_VM_ONLY",
        "trigger": "REMEDIATION_TRIGGER_MANUAL",
        "state": "PENDING_APPROVAL",
        "reason": "<string>",
        "active_health_check_run_id": "<string>",
        "passive_health_check_event_id": "<string>",
        "requested_by": "<string>",
        "create_time": "2023-11-07T05:31:56Z",
        "reviewed_by": "<string>",
        "review_time": "2023-11-07T05:31:56Z",
        "review_comment": "<string>",
        "start_time": "2023-11-07T05:31:56Z",
        "end_time": "2023-11-07T05:31:56Z",
        "error_message": "<string>",
        "update_time": "2023-11-07T05:31:56Z",
        "instance_name": "<string>",
        "linked_alerts": [
          {
            "passive_health_check_alert_id": "<string>",
            "cluster_id": "<string>",
            "target_vm": "<string>",
            "alert_name": "<string>",
            "severity": "PHC_SEVERITY_INFO",
            "annotations": {},
            "started_at": "2023-11-07T05:31:56Z",
            "annotation": {
              "title": "<string>",
              "description": "<string>",
              "summary_line": "<string>",
              "xid": {
                "events": [
                  {
                    "xid_code": "<string>",
                    "mnemonic": "<string>",
                    "count": 123
                  }
                ]
              },
              "slurm_node_unavailable": {
                "reason": "<string>"
              }
            },
            "instance_id": "<string>",
            "resolved_at": "2023-11-07T05:31:56Z",
            "node_remediation_intent_id": "<string>"
          }
        ]
      },
      "slurm_worker_hostname": "<string>",
      "marked_for_deletion": true,
      "public_ipv4": "<string>",
      "ib_hca_type": "<string>",
      "ib_hca_count": 123,
      "nvswitch_count": 123,
      "nvswitch_type": "<string>",
      "ephemeral_storage": "<string>",
      "auto_remediation_enabled": true,
      "deleted_at": "2023-11-07T05:31:56Z"
    }
  ],
  "kube_config": "<string>",
  "num_gpus": 123,
  "cuda_version": "<string>",
  "nvidia_driver_version": "<string>",
  "project_id": "<string>",
  "num_cpu_workers": 123,
  "phase_transitions": [
    {
      "phase": "CLUSTER_PHASE_QUEUED",
      "transition_time": "2023-11-07T05:31:56Z"
    }
  ],
  "desired_preemptible_gpus": 123,
  "allocated_preemptible_gpus": 123,
  "billing_type": "RESERVED",
  "add_ons": [
    {
      "name": "<string>",
      "add_on_type": "<string>",
      "config": {
        "dashboard": {
          "enabled": true
        },
        "ingress": {
          "enabled": true
        },
        "torchpass": {
          "enabled": true
        },
        "slurm_web": {
          "enabled": true
        },
        "headlamp": {
          "enabled": true
        }
      },
      "state": {
        "dashboard": {},
        "ingress": {},
        "torchpass": {},
        "slurm_web": {},
        "headlamp": {}
      }
    }
  ],
  "num_capacity_pool_gpus": 123,
  "num_reserved_gpus": 123,
  "duration_hours": 123,
  "slurm_shm_size_gib": 123,
  "capacity_pool_id": "<string>",
  "reservation_start_time": "2023-11-07T05:31:56Z",
  "reservation_end_time": "2023-11-07T05:31:56Z",
  "install_traefik": true,
  "created_at": "2023-11-07T05:31:56Z",
  "oidc_config": {
    "issuer_url": "<string>",
    "client_id": "<string>",
    "username_claim": "<string>",
    "username_prefix": "<string>",
    "group_claim": "<string>",
    "group_prefix": "<string>",
    "ca_cert": "<string>"
  },
  "cluster_config": {
    "load_balancer": "NONE",
    "kubernetes_dashboard_enabled": true,
    "jumphost_enabled": true,
    "slurm_startup_scripts": {
      "worker_prolog": "<string>",
      "worker_epilog": "<string>",
      "controller_prolog": "<string>",
      "controller_epilog": "<string>",
      "login_init_script": "<string>",
      "nodeset_init_script": "<string>",
      "extra_slurm_conf": "<string>"
    },
    "ingress": {
      "enabled": true
    },
    "observability": {
      "enabled": true
    },
    "gpu_operator_version": "<string>",
    "network_operator_version": "<string>",
    "ssh_ca_enabled": true
  },
  "machine_cluster_id": "<string>",
  "first_ready_at": "2023-11-07T05:31:56Z",
  "is_in_substrate": true,
  "control_plane_ready": true,
  "ums_project_id": "<string>",
  "ums_org_id": "<string>",
  "os_image": "<string>",
  "nvidia_driver_version_id": "<string>",
  "deleted_gpu_worker_nodes": [
    {
      "node_id": "<string>",
      "status": "<string>",
      "host_name": "<string>",
      "num_cpu_cores": 123,
      "num_gpus": 123,
      "memory_gib": 123,
      "networks": [
        "<string>"
      ],
      "phase_transitions": [
        {
          "phase": "NODE_PHASE_PENDING",
          "transition_time": "2023-11-07T05:31:56Z"
        }
      ],
      "instance_id": "<string>",
      "latest_remediation": {
        "id": "<string>",
        "cluster_id": "<string>",
        "instance_id": "<string>",
        "mode": "REMEDIATION_MODE_VM_ONLY",
        "trigger": "REMEDIATION_TRIGGER_MANUAL",
        "state": "PENDING_APPROVAL",
        "reason": "<string>",
        "active_health_check_run_id": "<string>",
        "passive_health_check_event_id": "<string>",
        "requested_by": "<string>",
        "create_time": "2023-11-07T05:31:56Z",
        "reviewed_by": "<string>",
        "review_time": "2023-11-07T05:31:56Z",
        "review_comment": "<string>",
        "start_time": "2023-11-07T05:31:56Z",
        "end_time": "2023-11-07T05:31:56Z",
        "error_message": "<string>",
        "update_time": "2023-11-07T05:31:56Z",
        "instance_name": "<string>",
        "linked_alerts": [
          {
            "passive_health_check_alert_id": "<string>",
            "cluster_id": "<string>",
            "target_vm": "<string>",
            "alert_name": "<string>",
            "severity": "PHC_SEVERITY_INFO",
            "annotations": {},
            "started_at": "2023-11-07T05:31:56Z",
            "annotation": {
              "title": "<string>",
              "description": "<string>",
              "summary_line": "<string>",
              "xid": {
                "events": [
                  {
                    "xid_code": "<string>",
                    "mnemonic": "<string>",
                    "count": 123
                  }
                ]
              },
              "slurm_node_unavailable": {
                "reason": "<string>"
              }
            },
            "instance_id": "<string>",
            "resolved_at": "2023-11-07T05:31:56Z",
            "node_remediation_intent_id": "<string>"
          }
        ]
      },
      "slurm_worker_hostname": "<string>",
      "marked_for_deletion": true,
      "public_ipv4": "<string>",
      "ib_hca_type": "<string>",
      "ib_hca_count": 123,
      "nvswitch_count": 123,
      "nvswitch_type": "<string>",
      "ephemeral_storage": "<string>",
      "auto_remediation_enabled": true,
      "deleted_at": "2023-11-07T05:31:56Z"
    }
  ],
  "node_lifecycle_events": [
    {
      "node_id": "<string>",
      "reason": "<string>",
      "message": "<string>",
      "timestamp": "2023-11-07T05:31:56Z"
    }
  ]
}

Authorizations

Authorization · string · header · default: default · required

Bearer authentication header of the form Bearer <token>, where <token> is your auth token.

Path Parameters

cluster_id · string · required

The ID of the cluster to update

Body

application/json

cluster_type · enum<string>

Type of cluster to update.

Available options: KUBERNETES, SLURM

num_gpus · integer

Target GPU count for the cluster. When omitted, the server keeps the current GPU count from cluster metadata (use for config-only or decommission-time-only updates).

reservation_end_time · string<date-time>

Timestamp at which the cluster should be decommissioned. Only accepted for prepaid clusters.

cluster_config · object

Show child attributes

num_reserved_gpus · integer

Number of reserved GPUs to update to. This field is only applicable for clusters with RESERVED billing type.

num_preemptible_gpus · integer

Updated desired number of preemptible GPUs for the cluster. When omitted, the current value is preserved. Must be a multiple of 8.

num_capacity_pool_gpus · integer<int32>

Number of GPUs to draw from the cluster's capacity pool. Only valid for clusters created with a capacity_pool_id. Must be a multiple of 8 and not exceed num_gpus. When omitted, the current value is preserved.

add_ons · object[]

Add-ons to update on the cluster. Each entry identifies an existing add-on by name and provides the new external config to merge.

Show child attributes

Response

200 - application/json

OK

cluster_id · string · required

cluster_type · enum<string> · required

Type of cluster.

Available options: KUBERNETES, SLURM

region · string · required

gpu_type · enum<string> · required

Available options: H100_SXM, H200_SXM, RTX_6000_PCI, L40_PCIE, B200_SXM, H100_SXM_INF, B300_SXM

cluster_name · string · required

volumes · object[] · required

Show child attributes

status · enum<string> · required

Current status of the GPU cluster.

Available options: WaitingForControlPlaneNodes, WaitingForDataPlaneNodes, WaitingForSubnet, WaitingForSharedVolume, InstallingDrivers, RunningAcceptanceTests, Paused, OnDemandComputePaused, Ready, Degraded, Deleting

control_plane_nodes · object[] · required

Show child attributes

gpu_worker_nodes · object[] · required

Show child attributes

kube_config · string · required

num_gpus · integer · required

cuda_version · string · required

nvidia_driver_version · string · required

project_id · string · required

num_cpu_workers · integer · required

Number of CPU-only worker nodes in the cluster.

phase_transitions · object[] · required

Cluster-level phase transition history.

Show child attributes

desired_preemptible_gpus · integer · required

Customer's requested number of preemptible GPUs. Set on cluster create or update; persists until changed.

allocated_preemptible_gpus · integer · required

Actual number of preemptible GPUs currently allocated to the cluster. Updated asynchronously by the fulfillment and reclamation workers; may be less than desired_preemptible_gpus when capacity is constrained.

billing_type · enum<string> · required

Billing type for the cluster (RESERVED, ON_DEMAND, or SCHEDULED_CAPACITY).

Available options: RESERVED, ON_DEMAND, SCHEDULED_CAPACITY

add_ons · object[] · required

Enabled add-ons on this cluster. Only add-ons with enabled=true in their config are returned.

Show child attributes

num_capacity_pool_gpus · integer<int32> · required

Number of GPUs to draw from a capacity pool. A component of the overall num_gpus, alongside num_reserved_gpus.

num_reserved_gpus · integer<int32> · required

Number of prepaid reserved GPUs for this cluster. A component of the overall num_gpus, alongside num_capacity_pool_gpus.

duration_hours · integer

slurm_shm_size_gib · integer

capacity_pool_id · string

reservation_start_time · string<date-time>

reservation_end_time · string<date-time>

install_traefik · boolean

created_at · string<date-time>

oidc_config · object

Show child attributes

cluster_config · object

Show child attributes

machine_cluster_id · string

ID of the machine cluster backing this GPU cluster.

first_ready_at · string<date-time>

Timestamp when the cluster first reached the Ready phase.

is_in_substrate · boolean

Whether the cluster is managed inside a substrate environment.

control_plane_ready · boolean

Whether the control plane is currently ready.

ums_project_id · string

UMS project ID associated with this cluster.

ums_org_id · string

UMS organization ID associated with this cluster.

os_image · string

Data-volume image name for GPU worker nodes.

nvidia_driver_version_id · string

Internal NVIDIA version ID for this cluster's driver and CUDA combination.

deleted_gpu_worker_nodes · object[]

GPU worker nodes retained after they left the live data plane. These are separate from gpu_worker_nodes and must not be counted as live capacity.

Show child attributes

node_lifecycle_events · object[]

Recent node lifecycle events such as scale-up, scale-down, and preemption. Combine these with live and deleted node lists to render the cluster timeline.

Show child attributes