Get training job metrics

Get the metrics for a training job.

GET /v1/training_projects/{training_project_id}/jobs/{training_job_id}/metrics

cURL

cURL

curl --request GET \
--url https://api.baseten.co/v1/training_projects/{training_project_id}/jobs/{training_job_id}/metrics \
--header "Authorization: Bearer $BASETEN_API_KEY"

200

{
  "gpu_memory_usage_bytes": {},
  "gpu_utilization": {},
  "cpu_usage": [
    {
      "value": 123,
      "timestamp": "2023-11-07T05:31:56Z"
    }
  ],
  "cpu_memory_usage_bytes": [
    {
      "value": 123,
      "timestamp": "2023-11-07T05:31:56Z"
    }
  ],
  "ephemeral_storage": {
    "usage_bytes": [
      {
        "value": 123,
        "timestamp": "2023-11-07T05:31:56Z"
      }
    ],
    "utilization": [
      {
        "value": 123,
        "timestamp": "2023-11-07T05:31:56Z"
      }
    ]
  },
  "training_job": {
    "id": "<string>",
    "created_at": "2023-11-07T05:31:56Z",
    "current_status": "<string>",
    "instance_type": {
      "id": "<string>",
      "name": "<string>",
      "memory_limit_mib": 123,
      "millicpu_limit": 123,
      "gpu_count": 123,
      "gpu_type": "<string>",
      "gpu_memory_limit_mib": 123
    },
    "updated_at": "2023-11-07T05:31:56Z",
    "training_project_id": "<string>",
    "training_project": {
      "id": "<string>",
      "name": "<string>"
    },
    "error_message": "<string>",
    "name": "gpt-oss-job",
    "priority": 0,
    "availability_model": "dedicated",
    "user": {
      "email": "<string>"
    }
  },
  "cache": {
    "usage_bytes": [
      {
        "value": 123,
        "timestamp": "2023-11-07T05:31:56Z"
      }
    ],
    "utilization": [
      {
        "value": 123,
        "timestamp": "2023-11-07T05:31:56Z"
      }
    ]
  },
  "per_node_metrics": [
    {
      "node_id": "<string>",
      "metrics": {
        "gpu_memory_usage_bytes": {},
        "gpu_utilization": {},
        "cpu_usage": [
          {
            "value": 123,
            "timestamp": "2023-11-07T05:31:56Z"
          }
        ],
        "cpu_memory_usage_bytes": [
          {
            "value": 123,
            "timestamp": "2023-11-07T05:31:56Z"
          }
        ],
        "ephemeral_storage": {
          "usage_bytes": [
            {
              "value": 123,
              "timestamp": "2023-11-07T05:31:56Z"
            }
          ],
          "utilization": [
            {
              "value": 123,
              "timestamp": "2023-11-07T05:31:56Z"
            }
          ]
        }
      }
    }
  ]
}

Authorizations

Authorization · string · header · required

Send Authorization: Bearer <api_key>. The legacy Authorization: Api-Key <api_key> scheme is also accepted.

Path Parameters

training_project_id · string · required

training_job_id · string · required

Query Parameters

end_epoch_millis · integer | null

Epoch millis timestamp to end fetching metrics

start_epoch_millis · integer | null

Epoch millis timestamp to start fetching metrics.

step_seconds · integer | null

Resolution of the returned series, in seconds. When omitted, a step is derived from the time range so large windows return fewer points.

Response

200 - application/json

A response to fetch training job metrics. The outer list for each metric represents that metric across time.

gpu_memory_usage_bytes · Gpu Memory Usage Bytes · object · required

A map of GPU rank to memory usage for the training job. For multinode jobs, this is the memory usage of the leader unless specified otherwise.

Show child attributes

gpu_utilization · Gpu Utilization · object · required

A map of GPU rank to fractional GPU utilization. For multinode jobs, this is the GPU utilization of the leader unless specified otherwise.

Show child attributes

cpu_usage · TrainingJobMetricV1 · object[] · required

The CPU usage measured in cores. For multinode jobs, this is the CPU usage of the leader unless specified otherwise.

Show child attributes

cpu_memory_usage_bytes · TrainingJobMetricV1 · object[] · required

The CPU memory usage for the training job. For multinode jobs, this is the CPU memory usage of the leader unless specified otherwise.

Show child attributes

ephemeral_storage · StorageMetricsV1 · object · required

The storage usage for the ephemeral storage. For multinode jobs, this is the ephemeral storage usage of the leader unless specified otherwise.

Show child attributes

training_job · TrainingJobV1 · object · required

The training job.

Show child attributes

cache · StorageMetricsV1 · object | null · required

The storage usage for the read-write cache.

Show child attributes

per_node_metrics · TrainingJobNodeMetricsV1 · object[] · required

The metrics for each node in the training job.

Show child attributes