Edit

Share via

Endpoint - Update ML Model Endpoint

Service:
MLModel
API Version:
v1

Updates the default version of the specified model.

Permissions

The caller must have write permission on the MLModel.

Required Delegated Scopes

MLModel.ReadWrite.All or Item.ReadWrite.All

Microsoft Entra supported identities

This API supports the Microsoft identities listed in this section.

Identity Support
User Yes
Service principal and Managed identities Yes

Interface

PATCH https://api.fabric.microsoft.com/v1/workspaces/{workspaceId}/mlmodels/{modelId}/endpoint

URI Parameters

Name In Required Type Description
modelId
 path True

string (uuid)

The machine learning model ID.
workspaceId
 path True

string (uuid)

The workspace ID.

Request Body

Name Type Description
defaultVersionAssignmentBehavior

EndpointDefaultVersionConfigurationPolicy

The default version assignment behavior of a given machine learning model endpoint. Additional EndpointDefaultVersionConfigurationPolicy types may be added over time.
defaultVersionName

string

Default machine learning model endpoint version name.

Responses

Name Type Description
200 OK

MLModelEndpoint

Endpoint updated.
Other Status Codes

ErrorResponse

Common error codes:

  • NotFound - Version not found.

  • UnknownError - An error occurred.

  • Conflict - An operation is already is proress.

  • BadInput - User Input is invalid.

Examples

Example of Update ML Model Endpoint
Example of Update ML Model Endpoint with an Activation Failure

Example of Update ML Model Endpoint

Sample request

PATCH https://api.fabric.microsoft.com/v1/workspaces/cf5fef71-f7f3-43f3-ae7f-3c427922bef1/mlmodels/7e92a6fc-3ef5-40bf-96df-ddae3fcde313/endpoint

{
  "defaultVersionAssignmentBehavior": "StaticallyConfigured",
  "defaultVersionName": "1"
}

Sample response

Status code:
200 
{
  "defaultVersionAssignmentBehavior": "StaticallyConfigured",
  "defaultVersionName": "1",
  "defaultVersionInfo": {
    "versionName": "1",
    "status": "Running",
    "inputSignature": [
      {
        "type": "double",
        "name": "age",
        "required": true
      },
      {
        "type": "double",
        "name": "sex",
        "required": true
      },
      {
        "type": "double",
        "name": "bmi",
        "required": true
      },
      {
        "type": "double",
        "name": "bp",
        "required": true
      },
      {
        "type": "double",
        "name": "s1",
        "required": true
      },
      {
        "type": "double",
        "name": "s2",
        "required": true
      },
      {
        "type": "double",
        "name": "s3",
        "required": true
      },
      {
        "type": "double",
        "name": "s4",
        "required": true
      },
      {
        "type": "double",
        "name": "s5",
        "required": true
      },
      {
        "type": "double",
        "name": "s6",
        "required": true
      }
    ],
    "outputSignature": [
      {
        "type": "double",
        "name": "target",
        "required": true
      }
    ],
    "scaleRule": "AlwaysOn"
  }
}

Example of Update ML Model Endpoint with an Activation Failure

Sample request

PATCH https://api.fabric.microsoft.com/v1/workspaces/cf5fef71-f7f3-43f3-ae7f-3c427922bef1/mlmodels/7e92a6fc-3ef5-40bf-96df-ddae3fcde313/endpoint

{
  "defaultVersionAssignmentBehavior": "StaticallyConfigured",
  "defaultVersionName": "1"
}

Sample response

Status code:
200 
{
  "defaultVersionAssignmentBehavior": "StaticallyConfigured",
  "defaultVersionName": "1",
  "defaultVersionInfo": {
    "versionName": "1",
    "status": "Failed",
    "failureDetails": {
      "errorCode": "InvalidInput",
      "message": "Tensor-based models are not supported yet."
    }
  }
}

Definitions

Name Description
EndpointDefaultVersionConfigurationPolicy

The default version assignment behavior of a given machine learning model endpoint. Additional EndpointDefaultVersionConfigurationPolicy types may be added over time.
ErrorRelatedResource

The error related resource details object.
ErrorResponse

The error response.
ErrorResponseDetails

The error response details.
FailureDetails

Activation Failure details.
MLModelDataSchema

Machine learning model data schema.
MLModelEndpoint

A machine learning model endpoint object.
MLModelEndpointVersionInfo

Machine Learning Model Endpoint version information.
ModelEndpointVersionStatus

The status of a given machine learning model endpoint version. Additional ModelEndpointVersionStatus types may be added over time.
ScaleRule

Machine learning model endpoint scale rule. Additional ScaleRule types may be added over time.
UpdateMLModelEndpointRequest

Machine learning model endpoint request body to update Model Endpoint properties.

EndpointDefaultVersionConfigurationPolicy

Enumeration

The default version assignment behavior of a given machine learning model endpoint. Additional EndpointDefaultVersionConfigurationPolicy types may be added over time.

Value Description
StaticallyConfigured

The default version is statically configured.
NotConfigured

No default version has been configured.

ErrorRelatedResource

Object

The error related resource details object.

Name Type Description
resourceId

string

The resource ID that's involved in the error.
resourceType

string

The type of the resource that's involved in the error.

ErrorResponse

Object

The error response.

Name Type Description
errorCode

string

A specific identifier that provides information about an error condition, allowing for standardized communication between our service and its users.
message

string

A human readable representation of the error.
moreDetails

ErrorResponseDetails[]

List of additional error details.
relatedResource

ErrorRelatedResource

The error related resource details.
requestId

string

ID of the request associated with the error.

ErrorResponseDetails

Object

The error response details.

Name Type Description
errorCode

string

A specific identifier that provides information about an error condition, allowing for standardized communication between our service and its users.
message

string

A human readable representation of the error.
relatedResource

ErrorRelatedResource

The error related resource details.

FailureDetails

Object

Activation Failure details.

Name Type Description
errorCode

string

A specific identifier that provides information about an error condition, allowing for standardized communication between our service and its users.
message

string

A human readable representation of the error.
relatedResource

ErrorRelatedResource

The error related resource details.

MLModelDataSchema

Object

Machine learning model data schema.

Name Type Description
name

string

The name of the signature.
required

boolean

The boolean value to indicate if the signature is required or no. True - Required, False - Not Required.
type

string

The type of the signature.

MLModelEndpoint

Object

A machine learning model endpoint object.

Name Type Description
defaultVersionAssignmentBehavior

EndpointDefaultVersionConfigurationPolicy

The default version assignment behavior of a given machine learning model endpoint. Additional EndpointDefaultVersionConfigurationPolicy types may be added over time.
defaultVersionInfo

MLModelEndpointVersionInfo

Machine Learning Model Endpoint version information.
defaultVersionName

string

Default machine learning model endpoint version name.

MLModelEndpointVersionInfo

Object

Machine Learning Model Endpoint version information.

Name Type Description
failureDetails

FailureDetails

Activation Failure details.
inputSignature

MLModelDataSchema[]

The input signature for this version of the machine learning model.
outputSignature

MLModelDataSchema[]

The output signature for this version of the machine learning model.
scaleRule

ScaleRule

Machine learning model endpoint scale rule. Additional ScaleRule types may be added over time.
status

ModelEndpointVersionStatus

The status of a given machine learning model endpoint version. Additional ModelEndpointVersionStatus types may be added over time.
versionName

string

The machine learning model version name

ModelEndpointVersionStatus

Enumeration

The status of a given machine learning model endpoint version. Additional ModelEndpointVersionStatus types may be added over time.

Value Description
Deactivated

All resources are deactivated.
Activating

Machine learning model endpoint activation is under progress.
Deactivating

Machine learning model endpoint deactivation is in progress.
Active

Machine learning model endpoint is available to serve score requests.
Failed

Machine learning model endpoint is in failed state.

ScaleRule

Enumeration

Machine learning model endpoint scale rule. Additional ScaleRule types may be added over time.

Value Description
AlwaysOn

Machine learning model endpoint should be AlwaysOn.
AllowScaleToZero

Machine learning model endpoint can be scaled to zero after a period of inactivity.

UpdateMLModelEndpointRequest

Object

Machine learning model endpoint request body to update Model Endpoint properties.

Name Type Description
defaultVersionAssignmentBehavior

EndpointDefaultVersionConfigurationPolicy

The default version assignment behavior of a given machine learning model endpoint. Additional EndpointDefaultVersionConfigurationPolicy types may be added over time.
defaultVersionName

string

Default machine learning model endpoint version name.