Endpoint - Update ML Model Endpoint Version

Service:

MLModel

API Version:

v1

Update machine learning model endpoint version configuration.

Permissions

The caller must have write permission on the MLModel.

Required Delegated Scopes

MLModel.ReadWrite.All or Item.ReadWrite.All

Limitations

• Only machine learning model endpoint versions which are enabled can be updated.

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/versions/{name}

URI Parameters

Name	In	Required	Type	Description
modelId	path	True	string (uuid)	The machine learning model ID.
name	path	True	string	The MLModel version name.
workspaceId	path	True	string (uuid)	The workspace ID.

Request Body

Name	Type	Description
scaleRule	ScaleRule	Machine learning model endpoint scale rule. Additional ScaleRule types may be added over time.

Responses

Name	Type	Description
200 OK	MLModelEndpointVersionInfo	Endpoint version configuration details.
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 Version

Example of Update ML Model Endpoint Version with an Activation Failure

Example of Update ML Model Endpoint Version

Sample request

HTTP

PATCH https://api.fabric.microsoft.com/v1/workspaces/cf5fef71-f7f3-43f3-ae7f-3c427922bef1/mlmodels/7e92a6fc-3ef5-40bf-96df-ddae3fcde313/endpoint/versions/1

{
  "scaleRule": "AllowScaleToZero"
}

Sample response

Status code:

200

{
  "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": "AllowScaleToZero"
}

Example of Update ML Model Endpoint Version with an Activation Failure

Sample request

HTTP

PATCH https://api.fabric.microsoft.com/v1/workspaces/cf5fef71-f7f3-43f3-ae7f-3c427922bef1/mlmodels/7e92a6fc-3ef5-40bf-96df-ddae3fcde313/endpoint/versions/1

{
  "scaleRule": "AllowScaleToZero"
}

Sample response

Status code:

200

{
  "versionName": "1",
  "status": "Failed",
  "failureDetails": {
    "errorCode": "InvalidInput",
    "message": "Tensor-based models are not supported yet."
  }
}

Definitions

Name	Description
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.
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.
UpdateMLModelEndpointVersionRequest	Machine learning model endpoint version configuration info.

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.

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.

UpdateMLModelEndpointVersionRequest

Object

Machine learning model endpoint version configuration info.

Name	Type	Description
scaleRule	ScaleRule	Machine learning model endpoint scale rule. Additional ScaleRule types may be added over time.