Azure Machine Learning Model Management Account API reference

For information about setting up the deployment environment, see Model Management account setup.

The Azure Machine Learning Model Management Account API implements the following operations:

  • Model registration
  • Manifest creation
  • Docker image creation
  • Web service creation

You can use this image to create a web service either locally or on a remote Azure Container Service cluster or another Docker-supported environment of your choice.

Prerequisites

Make sure you have gone through the installation steps in the Install and create Quickstart document.

The following are required before you proceed:

  1. Model Management account provisioning
  2. Environment creation for deploying and managing models
  3. A Machine Learning model

Azure AD token

When you're using Azure CLI, log in by using az login. The CLI uses your Azure Active Directory (Azure AD) token from the .azure file. If you want to use the APIs, you have the following options.

Acquire the Azure AD token manually

You can use az login and get the latest token from the .azure file on your home directory.

Acquire the Azure AD token programmatically
az ad sp create-for-rbac --scopes /subscriptions/<SubscriptionId>/resourcegroups/<ResourceGroupName> --role Contributor --years <length of time> --name <MyservicePrincipalContributor>

After you create the service principal, save the output. Now you can use that to get a token from Azure AD:

 private static async Task<string> AcquireTokenAsync(string clientId, string password, string authority, string resource)
{
        var creds = new ClientCredential(clientId, password);
        var context = new AuthenticationContext(authority);
        var token = await context.AcquireTokenAsync(resource, creds).ConfigureAwait(false);
        return token.AccessToken;
}

The token is put in an authorization header for API calls.

Register a model

The model registration step registers your Machine Learning model with the Azure Model Management account that you created. This registration enables tracking the models and their versions that are assigned at the time of registration. The user provides the name of the model. Subsequent registration of models under the same name generates a new version and ID.

Request

Method Request URI
POST /api/subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/accounts/{accountName}/models

Description

Registers a model.

Parameters

Name Located in Description Required Schema
subscriptionId path Azure subscription ID. Yes string
resourceGroupName path Name of the resource group in which the Model Management account is located. Yes string
accountName path Name of the Model Management account. Yes string
api-version query Version of the Microsoft.Machine.Learning resource provider API to use. Yes string
Authorization header Authorization token. It should be something like "Bearer XXXXXX." Yes string
model body Payload that is used to register a model. Yes Model

Responses

Code Description Schema
200 OK. The model registration succeeded. Model
default Error response that describes why the operation failed. ErrorResponse

Query the list of models in an account

Request

Method Request URI
GET /api/subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/accounts/{accountName}/models

Description

Queries the list of models in an account. You can filter the result list by tag and name. If no filter is passed, the query lists all the models in the account. The returned list is paginated, and the count of items in each page is an optional parameter.

Parameters

Name Located in Description Required Schema
subscriptionId path Azure subscription ID. Yes string
resourceGroupName path Name of the resource group in which the Model Management account is located. Yes string
accountName path Name of the Model Management account. Yes string
api-version query Version of the Microsoft.Machine.Learning resource provider API to use. Yes string
Authorization header Authorization token. It should be something like "Bearer XXXXXX." Yes string
name query Object name. No string
tag query Model tag. No string
count query Number of items to retrieve in a page. No string
$skipToken query Continuation token to retrieve the next page. No string

Responses

Code Description Schema
200 Success. PaginatedModelList
default Error response that describes why the operation failed. ErrorResponse

Get model details

Request

Method Request URI
GET /api/subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/accounts/{accountName}/models/{id}

Description

Gets a model by ID.

Parameters

Name Located in Description Required Schema
subscriptionId path Azure subscription ID. Yes string
resourceGroupName path Name of the resource group in which the Model Management account is located. Yes string
accountName path Name of the Model Management account. Yes string
id path Object ID. Yes string
api-version query Version of the Microsoft.Machine.Learning resource provider API to use. Yes string
Authorization header Authorization token. It should be something like "Bearer XXXXXX." Yes string

Responses

Code Description Schema
200 Success. Model
default Error response that describes why the operation failed. ErrorResponse

Register a manifest with the registered model and all dependencies

Request

Method Request URI
POST /api/subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/accounts/{accountName}/manifests

Description

Registers a manifest with the registered model and all its dependencies.

Parameters

Name Located in Description Required Schema
subscriptionId path Azure subscription ID. Yes string
resourceGroupName path Name of the resource group in which the Model Management account is located. Yes string
accountName path Name of the Model Management account. Yes string
api-version query Version of the Microsoft.Machine.Learning resource provider API to use. Yes string
Authorization header Authorization token. It should be something like "Bearer XXXXXX." Yes string
manifestRequest body Payload that is used to register a manifest. Yes Manifest

Responses

Code Description Schema
200 Manifest registration was successful. Manifest
default Error response that describes why the operation failed. ErrorResponse

Query the list of manifests in an account

Request

Method Request URI
GET /api/subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/accounts/{accountName}/manifests

Description

Queries the list of manifests in an account. You can filter the result list by model ID and manifest name. If no filter is passed, the query lists all the manifests in the account. The returned list is paginated, and the count of items in each page is an optional parameter.

Parameters

Name Located in Description Required Schema
subscriptionId path Azure subscription ID. Yes string
resourceGroupName path Name of the resource group in which the Model Management account is located. Yes string
accountName path Name of the Model Management account. Yes string
api-version query Version of the Microsoft.Machine.Learning resource provider API to use. Yes string
Authorization header Authorization token. It should be something like "Bearer XXXXXX." Yes string
modelId query Model ID. No string
manifestName query Manifest name. No string
count query Number of items to retrieve in a page. No string
$skipToken query Continuation token to retrieve the next page. No string

Responses

Code Description Schema
200 Success. PaginatedManifestList
default Error response that describes why the operation failed. ErrorResponse

Get manifest details

Request

Method Request URI
GET /api/subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/accounts/{accountName}/manifests/{id}

Description

Gets the manifest by ID.

Parameters

Name Located in Description Required Schema
subscriptionId path Azure subscription ID. Yes string
resourceGroupName path Name of the resource group in which the Model Management account is located. Yes string
accountName path Name of the Model Management account. Yes string
id path Object ID. Yes string
api-version query Version of the Microsoft.Machine.Learning resource provider API to use. Yes string
Authorization header Authorization token. It should be something like "Bearer XXXXXX." Yes string

Responses

Code Description Schema
200 Success. Manifest
default Error response that describes why the operation failed. ErrorResponse

Create an image

Request

Method Request URI
POST /api/subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/accounts/{accountName}/images

Description

Creates an image as a Docker image in Azure Container Registry.

Parameters

Name Located in Description Required Schema
subscriptionId path Azure subscription ID. Yes string
resourceGroupName path Name of the resource group in which the Model Management account is located. Yes string
accountName path Name of the Model Management account. Yes string
api-version query Version of the Microsoft.Machine.Learning resource provider API to use. Yes string
Authorization header Authorization token. It should be something like "Bearer XXXXXX." Yes string
imageRequest body Payload that is used to create an image. Yes ImageRequest

Responses

Code Description Headers Schema
202 Async operation location URL. A GET call will show you the status of the image creation task. Operation-Location
default Error response that describes why the operation failed. ErrorResponse

Query the list of images in an account

Request

Method Request URI
GET /api/subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/accounts/{accountName}/images

Description

Queries the list of images in an account. You can filter the result list by manifest ID and name. If no filter is passed, the query lists all the images in the account. The returned list is paginated, and the count of items in each page is an optional parameter.

Parameters

Name Located in Description Required Schema
subscriptionId path Azure subscription ID. Yes string
resourceGroupName path Name of the resource group in which the Model Management account is located. Yes string
accountName path Name of the Model Management account. Yes string
api-version query Version of the Microsoft.Machine.Learning resource provider API to use. Yes string
Authorization header Authorization token. It should be something like "Bearer XXXXXX." Yes string
manifestId query Manifest ID. No string
manifestName query Manifest name. No string
count query Number of items to retrieve in a page. No string
$skipToken query Continuation token to retrieve the next page. No string

Responses

Code Description Schema
200 Success. PaginatedImageList
default Error response that describes why the operation failed. ErrorResponse

Get image details

Request

Method Request URI
GET /api/subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/accounts/{accountName}/images/{id}

Description

Gets an image by ID.

Parameters

Name Located in Description Required Schema
subscriptionId path Azure subscription ID. Yes string
resourceGroupName path Name of the resource group in which the Model Management account is located. Yes string
accountName path Name of the Model Management account. Yes string
id path Image ID. Yes string
api-version query Version of the Microsoft.Machine.Learning resource provider API to use. Yes string
Authorization header Authorization token. It should be something like "Bearer XXXXXX." Yes string

Responses

Code Description Schema
200 Success. Image
default Error response that describes why the operation failed. ErrorResponse

Create a service

Request

Method Request URI
POST /api/subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/accounts/{accountName}/services

Description

Creates a service from an image.

Parameters

Name Located in Description Required Schema
subscriptionId path Azure subscription ID. Yes string
resourceGroupName path Name of the resource group in which the Model Management account is located. Yes string
accountName path Name of the Model Management account. Yes string
api-version query Version of the Microsoft.Machine.Learning resource provider API to use. Yes string
Authorization header Authorization token. It should be something like "Bearer XXXXXX." Yes string
serviceRequest body Payload that is used to create a service. Yes ServiceCreateRequest

Responses

Code Description Headers Schema
202 Async operation location URL. A GET call will show you the status of the service creation task. Operation-Location
409 A service with the specified name already exists.
default Error response that describes why the operation failed. ErrorResponse

Query the list of services in an account

Request

Method Request URI
GET /api/subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/accounts/{accountName}/services

Description

Queries the list of services in an account. You can filter the result list by model name/ID, manifest name/ID, image ID, service name, or Machine Learning compute resource ID. If no filter is passed, the query lists all services in the account. The returned list is paginated, and the count of items in each page is an optional parameter.

Parameters

Name Located in Description Required Schema
subscriptionId path Azure subscription ID. Yes string
resourceGroupName path Name of the resource group in which the Model Management account is located. Yes string
accountName path Name of the Model Management account. Yes string
api-version query Version of the Microsoft.Machine.Learning resource provider API to use. Yes string
Authorization header Authorization token. It should be something like "Bearer XXXXXX." Yes string
serviceName query Service name. No string
modelId query Model name. No string
modelName query Model ID. No string
manifestId query Manifest ID. No string
manifestName query Manifest name. No string
imageId query Image ID. No string
computeResourceId query Machine Learning compute resource ID. No string
count query Number of items to retrieve in a page. No string
$skipToken query Continuation token to retrieve the next page. No string

Responses

Code Description Schema
200 Success. PaginatedServiceList
default Error response that describes why the operation failed. ErrorResponse

Get service details

Request

Method Request URI
GET /api/subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/accounts/{accountName}/services/{id}

Description

Gets a service by ID.

Parameters

Name Located in Description Required Schema
subscriptionId path Azure subscription ID. Yes string
resourceGroupName path Name of the resource group in which the Model Management account is located. Yes string
accountName path Name of the Model Management account. Yes string
id path Object ID. Yes string
api-version query Version of the Microsoft.Machine.Learning resource provider API to use. Yes string
Authorization header Authorization token. It should be something like "Bearer XXXXXX." Yes string

Responses

Code Description Schema
200 Success. ServiceResponse
default Error response that describes why the operation failed. ErrorResponse

Update a service

Request

Method Request URI
PUT /api/subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/accounts/{accountName}/services/{id}

Description

Updates an existing service.

Parameters

Name Located in Description Required Schema
subscriptionId path Azure subscription ID. Yes string
resourceGroupName path Name of the resource group in which the Model Management account is located. Yes string
accountName path Name of the Model Management account. Yes string
id path Object ID. Yes string
api-version query Version of the Microsoft.Machine.Learning resource provider API to use. Yes string
Authorization header Authorization token. It should be something like "Bearer XXXXXX." Yes string
serviceUpdateRequest body Payload that is used to update an existing service. Yes ServiceUpdateRequest

Responses

Code Description Headers Schema
202 Async operation location URL. A GET call will show you the status of the update service task. Operation-Location
404 The service with the specified ID does not exist.
default Error response that describes why the operation failed. ErrorResponse

Delete a service

Request

Method Request URI
DELETE /api/subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/accounts/{accountName}/services/{id}

Description

Deletes a service.

Parameters

Name Located in Description Required Schema
subscriptionId path Azure subscription ID. Yes string
resourceGroupName path Name of the resource group in which the Model Management account is located. Yes string
accountName path Name of the Model Management account. Yes string
id path Object ID. Yes string
api-version query Version of the Microsoft.Machine.Learning resource provider API to use. Yes string
Authorization header Authorization token. It should be something like "Bearer XXXXXX." Yes string

Responses

Code Description Schema
200 Success.
204 The service with the specified ID does not exist.
default Error response that describes why the operation failed. ErrorResponse

Get service keys

Request

Method Request URI
GET /api/subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/accounts/{accountName}/services/{id}/keys

Description

Gets service keys.

Parameters

Name Located in Description Required Schema
subscriptionId path Azure subscription ID. Yes string
resourceGroupName path Name of the resource group in which the Model Management account is located. Yes string
accountName path Name of the Model Management account. Yes string
id path Service ID. Yes string
api-version query Version of the Microsoft.Machine.Learning resource provider API to use. Yes string
Authorization header Authorization token. It should be something like "Bearer XXXXXX." Yes string

Responses

Code Description Schema
200 Success. AuthKeys
default Error response that describes why the operation failed. ErrorResponse

Regenerate service keys

Request

Method Request URI
POST /api/subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/accounts/{accountName}/services/{id}/keys

Description

Regenerates service keys and returns them.

Parameters

Name Located in Description Required Schema
subscriptionId path Azure subscription ID. Yes string
resourceGroupName path Name of the resource group in which the Model Management account is located. Yes string
accountName path Name of the Model Management account. Yes string
id path Service ID. Yes string
api-version query Version of the Microsoft.Machine.Learning resource provider API to use. Yes string
Authorization header Authorization token. It should be something like "Bearer XXXXXX." Yes string
regenerateKeyRequest body Payload that is used to update an existing service. Yes ServiceRegenerateKeyRequest

Responses

Code Description Schema
200 Success. AuthKeys
default Error response that describes why the operation failed. ErrorResponse

Query the list of deployments in an account

Request

Method Request URI
GET /api/subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/accounts/{accountName}/deployments

Description

Queries the list of deployments in an account. You can filter the result list by service ID, which will return only the deployments that are created for the particular service. If no filter is passed, the query lists all the deployments in the account.

Parameters

Name Located in Description Required Schema
subscriptionId path Azure subscription ID. Yes string
resourceGroupName path Name of the resource group in which the Model Management account is located. Yes string
accountName path Name of the Model Management account. Yes string
api-version query Version of the Microsoft.Machine.Learning resource provider API to use. Yes string
Authorization header Authorization token. It should be something like "Bearer XXXXXX." Yes string
serviceId query Service ID. No string

Responses

Code Description Schema
200 Success. DeploymentList
default Error response that describes why the operation failed. ErrorResponse

Get deployment details

Request

Method Request URI
GET /api/subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/accounts/{accountName}/deployments/{id}

Description

Gets the deployment by ID.

Parameters

Name Located in Description Required Schema
subscriptionId path Azure subscription ID. Yes string
resourceGroupName path Name of the resource group in which the Model Management account is located. Yes string
accountName path Name of the Model Management account. Yes string
id path Deployment ID. Yes string
api-version query Version of the Microsoft.Machine.Learning resource provider API to use. Yes string
Authorization header Authorization token. It should be something like "Bearer XXXXXX." Yes string

Responses

Code Description Schema
200 Success. Deployment
default Error response that describes why the operation failed. ErrorResponse

Get operation details

Request

Method Request URI
GET /api/subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/accounts/{accountName}/operations/{id}

Description

Gets the async operation status by operation ID.

Parameters

Name Located in Description Required Schema
subscriptionId path Azure subscription ID. Yes string
resourceGroupName path Name of the resource group in which the Model Management account is located. Yes string
accountName path Name of the Model Management account. Yes string
id path Operation ID. Yes string
api-version query Version of the Microsoft.Machine.Learning resource provider API to use. Yes string
Authorization header Authorization token. It should be something like "Bearer XXXXXX." Yes string

Responses

Code Description Schema
200 Success. OperationStatus
default Error response that describes why the operation failed. ErrorResponse

Definitions

Asset

The asset object that will be needed during Docker image creation.

Name Description Schema
id
optional
Asset ID. string
mimeType
optional
MIME type of model content. For more information about MIME type, see the list of IANA media types. string
unpack
optional
Indicates where we need to unpack the content during Docker image creation. boolean
url
optional
Asset location URL. string

AsyncOperationState

The async operation state.

Type: enum (NotStarted, Running, Cancelled, Succeeded, Failed)

AsyncOperationStatus

The operation status.

Name Description Schema
createdTime
optional
read-only
Async operation creation time (UTC). string (date-time)
endTime
optional
read-only
Async operation end time (UTC). string (date-time)
error
optional
ErrorResponse
id
optional
Async operation ID. string
operationType
optional
Async operation type. enum (Image, Service)
resourceLocation
optional
Resource created or updated by the async operation. string
state
optional
AsyncOperationState

AuthKeys

The authentication keys for a service.

Name Description Schema
primaryKey
optional
Primary key. string
secondaryKey
optional
Secondary key. string

AutoScaler

Settings for the autoscaler.

Name Description Schema
autoscaleEnabled
optional
Enable or disable the autoscaler. boolean
maxReplicas
optional
Maximum number of pod replicas to scale up to.
Minimum value: 1
integer
minReplicas
optional
Minimum number of pod replicas to scale down to.
Minimum value: 0
integer
refreshPeriodInSeconds
optional
Refresh time for autoscaling trigger.
Minimum value: 1
integer
targetUtilization
optional
Utilization percentage that triggers autoscaling.
Minimum value: 0
Maximum value: 100
integer

ComputeResource

The Machine Learning compute resource.

Name Description Schema
id
optional
Resource ID. string
type
optional
Type of resource. enum (Cluster)

ContainerResourceReservation

Configuration to reserve resources for a container in the cluster.

Name Description Schema
cpu
optional
Specifies CPU reservation. Format for Kubernetes: see Meaning of CPU. string
memory
optional
Specifies memory reservation. Format for Kubernetes: see Meaning of memory. string

Deployment

An instance of an Azure Machine Learning deployment.

Name Description Schema
createdAt
optional
read-only
Deployment creation time (UTC). string (date-time)
expiredAt
optional
read-only
Deployment expired time (UTC). string (date-time)
id
optional
Deployment ID. string
imageId
optional
Image ID associated with this deployment. string
serviceName
optional
Service name. string
status
optional
Current deployment status. string

DeploymentList

An array of deployment objects.

Type: <Deployment> array

ErrorDetail

Model Management service error detail.

Name Description Schema
code
required
Error code. string
message
required
Error message. string

ErrorResponse

A Model Management service error object.

Name Description Schema
code
required
Error code. string
details
optional
Array of error detail objects. <ErrorDetail> array
message
required
Error message. string
statusCode
optional
HTTP status code. integer

Image

The Azure Machine Learning image.

Name Description Schema
computeResourceId
optional
ID of the environment created in the Machine Learning compute resource. string
createdTime
optional
Image creation time (UTC). string (date-time)
creationState
optional
AsyncOperationState
description
optional
Image description text. string
error
optional
ErrorResponse
id
optional
Image ID. string
imageBuildLogUri
optional
URI of the uploaded logs from the image build. string
imageLocation
optional
Azure Container Registry location string for the created image. string
imageType
optional
ImageType
manifest
optional
Manifest
name
optional
Image name. string
version
optional
Image version set by the Model Management service. integer

ImageRequest

A request to create an Azure Machine Learning image.

Name Description Schema
computeResourceId
required
ID of the environment created in the Machine Learning compute resource. string
description
optional
Image description text. string
imageType
required
ImageType
manifestId
required
ID of the manifest from which the image will be created. string
name
required
Image name. string

ImageType

Specifies the type of the image.

Type: enum (Docker)

Manifest

The Azure Machine Learning manifest.

Name Description Schema
assets
required
List of assets. <Asset> array
createdTime
optional
read-only
Manifest creation time (UTC). string (date-time)
description
optional
Manifest description text. string
driverProgram
required
Driver program of the manifest. string
id
optional
Manifest ID. string
modelIds
optional
List of model IDs of the registered models. The request will fail if any of the included models are not registered. array
modelType
optional
Specifies that the models are already registered with the Model Management service. enum (Registered)
name
required
Manifest name. string
targetRuntime
required
TargetRuntime
version
optional
read-only
Manifest version assigned by the Model Management service. integer
webserviceType
optional
Specifies the desired type of web service that will be created from the manifest. enum (Realtime)

Model

An instance of an Azure Machine Learning model.

Name Description Schema
createdAt
optional
read-only
Model creation time (UTC). string (date-time)
description
optional
Model description text. string
id
optional
read-only
Model ID. string
mimeType
required
MIME type of the model content. For more information about MIME type, see the list of IANA media types. string
name
required
Model name. string
tags
optional
Model tag list. array
unpack
optional
Indicates whether we need to unpack the model during Docker image creation. boolean
url
required
URL of the model. Usually we put a shared access signature URL here. string
version
optional
read-only
Model version assigned by the Model Management service. integer

ModelDataCollection

The model data collection information.

Name Description Schema
eventHubEnabled
optional
Enable an event hub for a service. boolean
storageEnabled
optional
Enable storage for a service. boolean

PaginatedImageList

A paginated list of images.

Name Description Schema
nextLink
optional
Continuation link (absolute URI) to the next page of results in the list. string
value
optional
Array of model objects. <Image> array

PaginatedManifestList

A paginated list of manifests.

Name Description Schema
nextLink
optional
Continuation link (absolute URI) to the next page of results in the list. string
value
optional
Array of manifest objects. <Manifest> array

PaginatedModelList

A paginated list of models.

Name Description Schema
nextLink
optional
Continuation link (absolute URI) to the next page of results in the list. string
value
optional
Array of model objects. <Model> array

PaginatedServiceList

A paginated list of services.

Name Description Schema
nextLink
optional
Continuation link (absolute URI) to the next page of results in the list. string
value
optional
Array of service objects. <ServiceResponse> array

ServiceCreateRequest

A request to create a service.

Name Description Schema
appInsightsEnabled
optional
Enable application insights for a service. boolean
autoScaler
optional
AutoScaler
computeResource
required
ComputeResource
containerResourceReservation
optional
ContainerResourceReservation
dataCollection
optional
ModelDataCollection
imageId
required
Image to create the service. string
maxConcurrentRequestsPerContainer
optional
Maximum number of concurrent requests.
Minimum value: 1
integer
name
required
Service name. string
numReplicas
optional
Number of pod replicas running at any time. Cannot be specified if Autoscaler is enabled.
Minimum value: 0
integer

ServiceRegenerateKeyRequest

A request to regenerate a key for a service.

Name Description Schema
keyType
optional
Specifies which key to regenerate. enum (Primary, Secondary)

ServiceResponse

The detailed status of the service.

Name Description Schema
createdAt
optional
Service creation time (UTC). string (date-time)
ID
optional
Service ID. string
image
optional
Image
manifest
optional
Manifest
models
optional
List of models. <Model> array
name
optional
Service name. string
scoringUri
optional
URI for scoring the service. string
state
optional
AsyncOperationState
updatedAt
optional
Last update time (UTC). string (date-time)
appInsightsEnabled
optional
Enable application insights for a service. boolean
autoScaler
optional
AutoScaler
computeResource
required
ComputeResource
containerResourceReservation
optional
ContainerResourceReservation
dataCollection
optional
ModelDataCollection
maxConcurrentRequestsPerContainer
optional
Maximum number of concurrent requests.
Minimum value: 1
integer
numReplicas
optional
Number of pod replicas running at any time. Cannot be specified if Autoscaler is enabled.
Minimum value: 0
integer
error
optional
ErrorResponse

ServiceUpdateRequest

A request to update a service.

Name Description Schema
appInsightsEnabled
optional
Enable application insights for a service. boolean
autoScaler
optional
AutoScaler
containerResourceReservation
optional
ContainerResourceReservation
dataCollection
optional
ModelDataCollection
imageId
optional
Image to create the service. string
maxConcurrentRequestsPerContainer
optional
Maximum number of concurrent requests.
Minimum value: 1
integer
numReplicas
optional
Number of pod replicas running at any time. Cannot be specified if Autoscaler is enabled.
Minimum value: 0
integer

TargetRuntime

The type of the target runtime.

Name Description Schema
properties
required
<string, string> map
runtimeType
required
Specifies the runtime. enum (SparkPython, Python)