Get nodes overview and states | NVIDIA Fleet Intelligence Documentation

{
  "hasMore": false,
  "nodes": [
    {
      "agentStatus": "Online",
      "agentVersion": "1.2.3",
      "computeZone": "compute-zone-1",
      "firmwareCheck": "Passed",
      "gpuCount": 4,
      "gpuDriverVersion": "550.120",
      "gpuFirmwareVersions": [
        {
          "firmwareVersion": "90.17.A9.00.0B",
          "gpuIndex": "0"
        }
      ],
      "gpuType": "NVIDIA-T400-4GB",
      "healthStatus": "Healthy",
      "hostname": "node-001",
      "integrityCheck": "Verified",
      "integrityCheckExtraInfo": {
        "errorDetails": "SDK error: code=13, message=nonce refresh failed",
        "gpuEvidencePresent": true,
        "gpuResults": {},
        "nrasCallResult": "NRAS attestation failed: GPU-0: NONCE_NOT_MATCHING",
        "overallAttestationResult": false
      },
      "integrityCheckReason": "Attestation verification unavailable: SDK error",
      "kernelVersion": "6.8.0-49-generic",
      "lastIntegrityCheckTS": "2024-01-15T10:30:00Z",
      "lastUpdatedTS": "2024-01-15T10:30:00Z",
      "nodeGroup": "node-group-1",
      "nodeUUID": "node-123",
      "privateIP": "10.0.0.100",
      "publicIP": "203.0.113.100"
    }
  ],
  "page": 0,
  "pageSize": 10,
  "total": 100
}

Get list of nodes with current state and overview. Use view=detail (default) for full response or view=basic for a lightweight nodeUUID + hostname list (see models.BasicNodesResponse). All filter parameters are optional arrays - empty means no filter applied.

Query parameters

viewenumOptionalDefaults to detail

View mode: ‘detail’ for full response (default) or ‘basic’ for nodeUUID + hostname only (lightweight query, does not support filtering or sorting by healthStatuses, integrityChecks, firmwareChecks, or agentStatuses).

Allowed values:

pageintegerOptionalDefaults to 0

Page number

pageSizeintegerOptionalDefaults to 10

Page size

nodeUUIDslist of stringsOptional

Filter by specific node UUIDs. Empty means all nodes

healthStatuseslist of enumsOptional

Filter by health status values

Allowed values:

computeZoneIdslist of stringsOptional

Filter by computezone IDs. Empty means all computezones

computeZoneNameslist of stringsOptional

Filter by computezone names (partial match). Empty means all computezones

nodeGroupIdslist of stringsOptional

Filter by nodegroup IDs. Empty means all nodegroups

nodeGroupNameslist of stringsOptional

Filter by nodegroup names (partial match). Empty means all nodegroups

gpuTypeslist of stringsOptional

Filter by GPU type values. Empty means all GPU types

gpuCountslist of integersOptional

Filter by GPU count values. Empty means all GPU counts

publicIPslist of stringsOptional

Filter by public IP addresses. Empty means all public IPs

privateIPslist of stringsOptional

Filter by private IP addresses. Empty means all private IPs

integrityCheckslist of enumsOptional

Filter by integrity check status

firmwareCheckslist of enumsOptional

Filter by firmware check status

Allowed values:

agentStatuseslist of enumsOptional

Filter by agent status values

Allowed values:

hostnamestringOptional

Filter by hostname (case-insensitive partial match)

sortByenumOptionalDefaults to hostname

Field to sort by. gpuFirmwareVersions pins nodes whose GPUs have mismatched vbios versions to the top, then sorts the remaining (uniform) nodes by that common vbios in the requested order.

orderenumOptionalDefaults to asc

Sort order (asc, desc)

Allowed values:

Response

Basic view response (view=basic)

models.NodesResponseobject

models.BasicNodesResponseobject

Errors

400

Bad Request Error

401

Unauthorized Error

403

Forbidden Error

499

Client Closed Request Error

500

Internal Server Error

503

Service Unavailable Error