Microsoft.MachineLearningServices workspaces/batchEndpoints/deployments 2024-04-01-preview

Article
11/21/2024

Bicep resource definition

The workspaces/batchEndpoints/deployments resource type can be deployed with operations that target:

Resource groups - See resource group deployment commands

For a list of changed properties in each API version, see change log.

Resource format

To create a Microsoft.MachineLearningServices/workspaces/batchEndpoints/deployments resource, add the following Bicep to your template.

resource symbolicname 'Microsoft.MachineLearningServices/workspaces/batchEndpoints/deployments@2024-04-01-preview' = {
  identity: {
    type: 'string'
    userAssignedIdentities: {
      {customized property}: {}
    }
  }
  kind: 'string'
  location: 'string'
  name: 'string'
  properties: {
    codeConfiguration: {
      codeId: 'string'
      scoringScript: 'string'
    }
    compute: 'string'
    deploymentConfiguration: {
      deploymentConfigurationType: 'string'
      // For remaining properties, see BatchDeploymentConfiguration objects
    }
    description: 'string'
    environmentId: 'string'
    environmentVariables: {
      {customized property}: 'string'
    }
    errorThreshold: int
    loggingLevel: 'string'
    maxConcurrencyPerInstance: int
    miniBatchSize: int
    model: {
      referenceType: 'string'
      // For remaining properties, see AssetReferenceBase objects
    }
    outputAction: 'string'
    outputFileName: 'string'
    properties: {
      {customized property}: 'string'
    }
    resources: {
      instanceCount: int
      instanceType: 'string'
      locations: [
        'string'
      ]
      maxInstanceCount: int
      properties: {
        {customized property}: any(Azure.Bicep.Types.Concrete.AnyType)
      }
    }
    retrySettings: {
      maxRetries: int
      timeout: 'string'
    }
  }
  sku: {
    capacity: int
    family: 'string'
    name: 'string'
    size: 'string'
    tier: 'string'
  }
  tags: {
    {customized property}: 'string'
  }
}

BatchDeploymentConfiguration objects

Set the deploymentConfigurationType property to specify the type of object.

For PipelineComponent, use:

{
  componentId: {
    assetId: 'string'
    referenceType: 'string'
  }
  deploymentConfigurationType: 'PipelineComponent'
  description: 'string'
  settings: {
    {customized property}: 'string'
  }
  tags: {
    {customized property}: 'string'
  }
}

AssetReferenceBase objects

Set the referenceType property to specify the type of object.

For DataPath, use:

{
  datastoreId: 'string'
  path: 'string'
  referenceType: 'DataPath'
}

For Id, use:

{
  assetId: 'string'
  referenceType: 'Id'
}

For OutputPath, use:

{
  jobId: 'string'
  path: 'string'
  referenceType: 'OutputPath'
}

Property values

AssetReferenceBase

Name	Description	Value
referenceType	Set to 'DataPath' for type DataPathAssetReference. Set to 'Id' for type IdAssetReference. Set to 'OutputPath' for type OutputPathAssetReference.	'DataPath' 'Id' 'OutputPath' (required)

BatchDeploymentConfiguration

Name	Description	Value
deploymentConfigurationType	Set to 'PipelineComponent' for type BatchPipelineComponentDeploymentConfiguration.	'PipelineComponent' (required)

BatchDeploymentProperties

Name	Description	Value
codeConfiguration	Code configuration for the endpoint deployment.	CodeConfiguration
compute	Compute target for batch inference operation.	string
deploymentConfiguration	Properties relevant to different deployment types.	BatchDeploymentConfiguration
description	Description of the endpoint deployment.	string
environmentId	ARM resource ID of the environment specification for the endpoint deployment.	string
environmentVariables	Environment variables configuration for the deployment.	EndpointDeploymentPropertiesBaseEnvironmentVariables
errorThreshold	Error threshold, if the error count for the entire input goes above this value, the batch inference will be aborted. Range is [-1, int.MaxValue]. For FileDataset, this value is the count of file failures. For TabularDataset, this value is the count of record failures. If set to -1 (the lower bound), all failures during batch inference will be ignored.	int
loggingLevel	Logging level for batch inference operation.	'Debug' 'Info' 'Warning'
maxConcurrencyPerInstance	Indicates maximum number of parallelism per instance.	int
miniBatchSize	Size of the mini-batch passed to each batch invocation. For FileDataset, this is the number of files per mini-batch. For TabularDataset, this is the size of the records in bytes, per mini-batch.	int
model	Reference to the model asset for the endpoint deployment.	AssetReferenceBase
outputAction	Indicates how the output will be organized.	'AppendRow' 'SummaryOnly'
outputFileName	Customized output file name for append_row output action.	string
properties	Property dictionary. Properties can be added, but not removed or altered.	EndpointDeploymentPropertiesBaseProperties
resources	Indicates compute configuration for the job. If not provided, will default to the defaults defined in ResourceConfiguration.	DeploymentResourceConfiguration
retrySettings	Retry Settings for the batch inference operation. If not provided, will default to the defaults defined in BatchRetrySettings.	BatchRetrySettings

BatchPipelineComponentDeploymentConfiguration

Name	Description	Value
componentId	The ARM id of the component to be run.	IdAssetReference
deploymentConfigurationType	[Required] The type of the deployment	'PipelineComponent' (required)
description	The description which will be applied to the job.	string
settings	Run-time settings for the pipeline job.	BatchPipelineComponentDeploymentConfigurationSettings
tags	The tags which will be applied to the job.	BatchPipelineComponentDeploymentConfigurationTags

BatchPipelineComponentDeploymentConfigurationSettings

Name	Description	Value

BatchPipelineComponentDeploymentConfigurationTags

Name	Description	Value

BatchRetrySettings

Name	Description	Value
maxRetries	Maximum retry count for a mini-batch	int
timeout	Invocation timeout for a mini-batch, in ISO 8601 format.	string

CodeConfiguration

Name	Description	Value
codeId	ARM resource ID of the code asset.	string
scoringScript	[Required] The script to execute on startup. eg. "score.py"	string Constraints: Min length = 1 Pattern = `[a-zA-Z0-9_]` (required)

DataPathAssetReference

Name	Description	Value
datastoreId	ARM resource ID of the datastore where the asset is located.	string
path	The path of the file/directory in the datastore.	string
referenceType	[Required] Specifies the type of asset reference.	'DataPath' (required)

DeploymentResourceConfiguration

Name	Description	Value
instanceCount	Optional number of instances or nodes used by the compute target.	int
instanceType	Optional type of VM used as supported by the compute target.	string
locations	Locations where the job can run.	string[]
maxInstanceCount	Optional max allowed number of instances or nodes to be used by the compute target. For use with elastic training, currently supported by PyTorch distribution type only.	int
properties	Additional properties bag.	ResourceConfigurationProperties

EndpointDeploymentPropertiesBaseEnvironmentVariables

Name	Description	Value

EndpointDeploymentPropertiesBaseProperties

Name	Description	Value

IdAssetReference

Name	Description	Value
assetId	[Required] ARM resource ID of the asset.	string Constraints: Min length = 1 Pattern = `[a-zA-Z0-9_]` (required)
referenceType	[Required] Specifies the type of asset reference.	'Id' (required)

IdAssetReference

Name	Description	Value
assetId	[Required] ARM resource ID of the asset.	string Constraints: Min length = 1 Pattern = `[a-zA-Z0-9_]` (required)
referenceType	[Required] Specifies the type of asset reference.	'DataPath' 'Id' 'OutputPath' (required)

ManagedServiceIdentity

Name	Description	Value
type	Type of managed service identity (where both SystemAssigned and UserAssigned types are allowed).	'None' 'SystemAssigned' 'SystemAssigned,UserAssigned' 'UserAssigned' (required)
userAssignedIdentities	The set of user assigned identities associated with the resource. The userAssignedIdentities dictionary keys will be ARM resource ids in the form: '/subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/providers/Microsoft.ManagedIdentity/userAssignedIdentities/{identityName}. The dictionary values can be empty objects ({}) in requests.	UserAssignedIdentities