GET
/
v1
/
models
/
{model_id}
/
metrics
curl --request GET \
  --url https://mango.sievedata.com/v1/models/{model_id}/metrics \
  --header 'X-API-Key: <api-key>'
{
  "model_requests": {
    "key": "model_requests",
    "name": "Model Requests",
    "description": "Number of requests sent to model",
    "type": "matrix",
    "value": [[1686336600, 0], [1686336720, 0]]
  },
  "queue_length": {
    "key": "queue_length",
    "name": "Queue Length",
    "description": "Number of requests in queue",
    "type": "matrix",
    "value": [[1686336600, 0], [1686336720, 0]]
  },
  "replica_count": {
    "key": "replica_count",
    "name": "Replica Count",
    "description": "Number of replicas",
    "type": "matrix",
    "value": [[1686336600, 0], [1686336720, 0]]
  }
}

This endpoint returns a JSON object that describes the metrics associated with a model given its ID.

​
Request

model_id
string
required

Sieve model ID

limit_min
number
default: "60"

The duration (in minutes) of historical metrics to retrieve, ending at the current time

​
Response

Responses are organized by key as seen below. The value key provides a list of timestamps and values for each metric.

model_requests
dict

The number of requests sent to a model at any given time.

queue_length
dict

The length of the queue for a model at any given time.

replica_count
dict

The number of replicas of the model that are up at any given time.

Was this page helpful?

List ModelsGet Model Requests