This API is used to query monitoring information about a single container of a job.

GET /v1/{project_id}/training-jobs/{job_id}/versions/{version_id}/pod/{pod_name}/metric-statistic

Table 1 describes the required parameters.

**Table 1** Parameter description
Parameter	Mandatory	Type	Description
project_id	Yes	String	Project ID
job_id	Yes	Long	ID of a training job
version_id	Yes	Long	Version ID of a training job
pod_name	Yes	String	Container name, which is the same as the job log name. For details about how to obtain the value, see Obtaining the Name of a Training Job Log File.

Table 2 describes the request parameters.

**Table 2** Parameter description
Parameter	Mandatory	Type	Description
metrics	No	String	Metrics to be queried. Separate metrics by commas (,), for example, CpuUsage,MemUsage. If this parameter is left blank, all metrics are queried. Options: CpuUsage MemUsage DiskReadRate DiskWriteRate RecvBytesRate SendBytesRate GpuUtil GpuMemUsage
statistic_type	No	String	Metric statistics method, indicating whether to collect metric statistics based on a single GPU. This parameter applies only to GPU metric statistics. all: Obtain the average value of the metric. each: Obtain the metric monitoring information about each GPU.

Table 3 describes the response parameters.

**Table 3** Parameter description
Parameter	Type	Description
error_message	String	Error message when the API call fails. This parameter is not included when the API call succeeds.
error_code	String	Error code when the API call fails. For details, see Error Codes. This parameter is not included when the API call succeeds.
metrics	JSON Array	Metric monitoring details. For details, see Table 4.
interval	Integer	Query interval, in minutes.

**Table 4** **metrics** data structure
Parameter	Type	Description
metric	String	Monitoring metrics
value	JSON Array	Sequence of the obtained metric value. The element is of the String type.

The following shows how to query the logs contained in log1.log of the job whose job_id is 10 and version_id is 10.

Sample request

GET    https://endpoint/v1/{project_id}/training-jobs/10/versions/10/pod/pod1/metric-statistic?metrics=gpuUtil

Successful sample response

{
  "metrics":
  [
    {
       "metric":"gpuUtil",
       "value":["1","22","33"]
     }
  ],
  "interval" : 1
}

Failed sample response

{
    "error_message": "Error string",
    "error_code": "ModelArts.0105"
}

For details about the status code, see Status Code.

Querying Monitoring Information About a Single Container of a Job