共用方式為


Microsoft.MachineLearningServices workspaces/onlineEndpoints/deployments 2023-04-01

Bicep 資源定義

workspaces/onlineEndpoints/deployments 資源類型可以使用目標作業來部署:

如需每個 API 版本中已變更屬性的清單,請參閱 變更記錄檔。

資源格式

若要建立 Microsoft.MachineLearningServices/workspaces/onlineEndpoints/deployments 資源,請將下列 Bicep 新增至您的範本。

resource symbolicname 'Microsoft.MachineLearningServices/workspaces/onlineEndpoints/deployments@2023-04-01' = {
  name: 'string'
  location: 'string'
  tags: {
    tagName1: 'tagValue1'
    tagName2: 'tagValue2'
  }
  sku: {
    capacity: int
    family: 'string'
    name: 'string'
    size: 'string'
    tier: 'string'
  }
  kind: 'string'
  parent: resourceSymbolicName
  identity: {
    type: 'string'
    userAssignedIdentities: {
      {customized property}: {}
    }
  }
  properties: {
    appInsightsEnabled: bool
    codeConfiguration: {
      codeId: 'string'
      scoringScript: 'string'
    }
    description: 'string'
    egressPublicNetworkAccess: 'string'
    environmentId: 'string'
    environmentVariables: {
      {customized property}: 'string'
      {customized property}: 'string'
    }
    instanceType: 'string'
    livenessProbe: {
      failureThreshold: int
      initialDelay: 'string'
      period: 'string'
      successThreshold: int
      timeout: 'string'
    }
    model: 'string'
    modelMountPath: 'string'
    properties: {
      {customized property}: 'string'
      {customized property}: 'string'
    }
    readinessProbe: {
      failureThreshold: int
      initialDelay: 'string'
      period: 'string'
      successThreshold: int
      timeout: 'string'
    }
    requestSettings: {
      maxConcurrentRequestsPerInstance: int
      maxQueueWait: 'string'
      requestTimeout: 'string'
    }
    scaleSettings: {
      scaleType: 'string'
      // For remaining properties, see OnlineScaleSettings objects
    }
    endpointComputeType: 'string'
    // For remaining properties, see OnlineDeploymentProperties objects
  }
}

OnlineDeploymentProperties 物件

設定 endpointComputeType 屬性以指定對象的類型。

針對 Kubernetes,請使用:

  endpointComputeType: 'Kubernetes'
  containerResourceRequirements: {
    containerResourceLimits: {
      cpu: 'string'
      gpu: 'string'
      memory: 'string'
    }
    containerResourceRequests: {
      cpu: 'string'
      gpu: 'string'
      memory: 'string'
    }
  }

針對 [受控],請使用:

  endpointComputeType: 'Managed'

OnlineScaleSettings 物件

設定 scaleType 屬性以指定物件的類型。

針對 [預設],請使用:

  scaleType: 'Default'

針對 TargetUtilization,請使用:

  scaleType: 'TargetUtilization'
  maxInstances: int
  minInstances: int
  pollingInterval: 'string'
  targetUtilizationPercentage: int

屬性值

workspaces/onlineEndpoints/deployments

名稱 描述
NAME 資源名稱

瞭解如何在 Bicep 中設定子資源的名稱和類型。
需要字串 ()
location 資源所在的地理位置 需要字串 ()
tags 資源標籤。 標記名稱和值的字典。 請參閱 範本中的標籤
sku 自動調整的 ARM 合約所需的 SKU 詳細數據。 Sku
kind 入口網站/工具/等用來轉譯相同類型資源的不同UX體驗的元數據。 字串
父系 (parent) 在 Bicep 中,您可以指定子資源的父資源。 只有在父資源外部宣告子資源時,才需要新增此屬性。

如需詳細資訊,請參閱 父資源外部的子資源
類型資源的符號名稱: onlineEndpoints
身分識別 受控服務識別 (系統指派和/或使用者指派的身分識別) ManagedServiceIdentity
properties [必要]實體的其他屬性。 OnlineDeploymentProperties (必要)

ManagedServiceIdentity

名稱 描述
類型 (允許 SystemAssigned 和 UserAssigned 類型的受控服務識別類型) 。 'None'
'SystemAssigned'
'SystemAssigned,UserAssigned'
'UserAssigned' (必要)
userAssignedIdentities 與資源相關聯的使用者指派身分識別集。 userAssignedIdentities 字典索引鍵的格式為 ARM 資源標識符:'/subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/providers/Microsoft.ManagedIdentity/userAssignedIdentities/{identityName}。 字典值可以是要求中 ({}) 的空白物件。 UserAssignedIdentities

UserAssignedIdentities

名稱 描述
{自定義屬性} UserAssignedIdentity

UserAssignedIdentity

此物件不包含部署期間要設定的任何屬性。 所有屬性都是 ReadOnly。

OnlineDeploymentProperties

名稱 描述
appInsightsEnabled 如果為 true,則啟用 Application Insights 記錄。 bool
codeConfiguration 端點部署的程式代碼組態。 CodeConfiguration
description 端點部署的描述。 字串
egressPublicNetworkAccess 如果已啟用,則允許輸出公用網路存取。 如果停用,這會建立安全的輸出。 預設值:已啟用。 'Disabled'
'Enabled'
environmentId 端點部署之環境規格的 ARM 資源識別碼或 AssetId。 字串
environmentVariables 部署的環境變數組態。 EndpointDeploymentPropertiesBaseEnvironmentVariables
instanceType 計算實例類型。 字串
livenessProbe 即時探查會定期監視容器的健康情況。 ProbeSettings
model 模型的 URI 路徑。 字串
modelMountPath 在自定義容器中掛接模型的路徑。 字串
properties 屬性字典。 您可以新增屬性,但無法移除或改變。 EndpointDeploymentPropertiesBaseProperties
readinessProbe 整備探查會驗證容器是否已準備好提供流量。 屬性和預設值與即時探查相同。 ProbeSettings
requestSettings 要求部署的設定。 OnlineRequestSettings
scaleSettings 部署的調整設定。
如果它是 Null 或未提供,則為
默認為 KubernetesOnlineDeployment 的 TargetUtilizationScaleSettings
和 至 ManagedOnlineDeployment 的 DefaultScaleSettings。
OnlineScaleSettings
endpointComputeType 設定物件類型 Kubernetes
受控 (必要)

CodeConfiguration

名稱 描述
codeId 程式代碼資產的 ARM 資源識別碼。 字串
scoringScript [必要]在啟動時執行的腳本。 例如 “score.py” 字串 (必要)

約束:
最小長度 = 1
模式 = [a-zA-Z0-9_]

EndpointDeploymentPropertiesBaseEnvironmentVariables

名稱 描述
{自定義屬性} 字串
{自定義屬性} 字串

ProbeSettings

名稱 描述
failureThreshold 傳回狀況不良狀態之前允許的失敗次數。 int
initialDelay ISO 8601 格式的第一個探查之前的延遲。 字串
Period ISO 8601 格式探查之間的時間長度。 字串
successThreshold 傳回狀況良好的狀態之前,成功探查的數目。 int
timeout ISO 8601 格式的探查逾時。 字串

EndpointDeploymentPropertiesBaseProperties

名稱 描述
{自定義屬性} 字串
{自定義屬性} 字串

OnlineRequestSettings

名稱 描述
maxConcurrentRequestsPerInstance 每個部署允許的每個節點並行要求數目上限。 預設值為 1。 int
maxQueueWait 要求會以 ISO 8601 格式保留在佇列中的最長時間量。
預設為 500 毫秒。
字串
requestTimeout ISO 8601 格式的評分逾時。
預設為 5000 毫秒。
字串

OnlineScaleSettings

名稱 描述
scaleType 設定物件類型 預設值
需要 TargetUtilization ()

DefaultScaleSettings

名稱 描述
scaleType [必要]部署調整演算法的類型 必要) 'Default' (

TargetUtilizationScaleSettings

名稱 描述
scaleType [必要]部署調整演算法的類型 'TargetUtilization' (必要)
maxInstances 部署可調整的實例數目上限。 配額將保留給max_instances。 int
minInstances 要一律存在的實例數目下限。 int
pollingInterval ISO 8691 格式的輪詢間隔。 僅支援有效位數為秒數的持續時間。 字串
targetUtilizationPercentage 自動調整程式的目標 CPU 使用量。 int

KubernetesOnlineDeployment

名稱 描述
endpointComputeType [必要]端點的計算類型。 必要) 'Kubernetes' (
containerResourceRequirements 容器的資源需求 (cpu 和記憶體) 。 ContainerResourceRequirements

ContainerResourceRequirements

名稱 描述
containerResourceLimits 容器資源限制資訊: ContainerResourceSettings
containerResourceRequests 容器資源要求資訊: ContainerResourceSettings

ContainerResourceSettings

名稱 描述
cpu 容器的 vCPU 要求/限制數目。 其他資訊:
https://kubernetes.io/docs/concepts/configuration/manage-compute-resources-container/
字串
gpu 容器的 Nvidia GPU 卡要求/限制數目。 其他資訊:
https://kubernetes.io/docs/concepts/configuration/manage-compute-resources-container/
字串
記憶體 容器的記憶體大小要求/限制。 其他資訊:
https://kubernetes.io/docs/concepts/configuration/manage-compute-resources-container/
字串

ManagedOnlineDeployment

名稱 描述
endpointComputeType [必要]端點的計算類型。 必要) 'Managed' (

SKU

名稱 描述
處理能力 如果 SKU 支援相應放大/縮小,則應該包含容量整數。 如果資源無法相應放大/縮小,可能會省略此專案。 int
family 如果服務有不同層代的硬體,則針對相同的SKU,則可以在這裡擷取。 字串
NAME SKU 的名稱。 例如 - P3。 通常是字母+數字代碼 字串 (必要)
{1}size{2} SKU 大小。 當名稱欄位是階層和其他一些值的組合時,這會是獨立程序代碼。 字串
tier 如果服務有多個層級,但 PUT 上不需要,則資源提供者必須實作此字段。 'Basic'
'Free'
'Premium'
'Standard'

ARM 範本資源定義

工作區/onlineEndpoints/deployments 資源類型可以使用目標作業來部署:

如需每個 API 版本中已變更屬性的清單,請參閱 變更記錄

資源格式

若要建立 Microsoft.MachineLearningServices/workspaces/onlineEndpoints/deployments 資源,請將下列 JSON 新增至您的範本。

{
  "type": "Microsoft.MachineLearningServices/workspaces/onlineEndpoints/deployments",
  "apiVersion": "2023-04-01",
  "name": "string",
  "location": "string",
  "tags": {
    "tagName1": "tagValue1",
    "tagName2": "tagValue2"
  },
  "sku": {
    "capacity": "int",
    "family": "string",
    "name": "string",
    "size": "string",
    "tier": "string"
  },
  "kind": "string",
  "identity": {
    "type": "string",
    "userAssignedIdentities": {
      "{customized property}": {}
    }
  },
  "properties": {
    "appInsightsEnabled": "bool",
    "codeConfiguration": {
      "codeId": "string",
      "scoringScript": "string"
    },
    "description": "string",
    "egressPublicNetworkAccess": "string",
    "environmentId": "string",
    "environmentVariables": {
      "{customized property}": "string",
      "{customized property}": "string"
    },
    "instanceType": "string",
    "livenessProbe": {
      "failureThreshold": "int",
      "initialDelay": "string",
      "period": "string",
      "successThreshold": "int",
      "timeout": "string"
    },
    "model": "string",
    "modelMountPath": "string",
    "properties": {
      "{customized property}": "string",
      "{customized property}": "string"
    },
    "readinessProbe": {
      "failureThreshold": "int",
      "initialDelay": "string",
      "period": "string",
      "successThreshold": "int",
      "timeout": "string"
    },
    "requestSettings": {
      "maxConcurrentRequestsPerInstance": "int",
      "maxQueueWait": "string",
      "requestTimeout": "string"
    },
    "scaleSettings": {
      "scaleType": "string"
      // For remaining properties, see OnlineScaleSettings objects
    },
    "endpointComputeType": "string"
    // For remaining properties, see OnlineDeploymentProperties objects
  }
}

OnlineDeploymentProperties 物件

設定 endpointComputeType 屬性以指定對象的類型。

針對 Kubernetes,請使用:

  "endpointComputeType": "Kubernetes",
  "containerResourceRequirements": {
    "containerResourceLimits": {
      "cpu": "string",
      "gpu": "string",
      "memory": "string"
    },
    "containerResourceRequests": {
      "cpu": "string",
      "gpu": "string",
      "memory": "string"
    }
  }

針對 Managed,請使用:

  "endpointComputeType": "Managed"

OnlineScaleSettings 物件

設定 scaleType 屬性以指定物件的類型。

針對 [預設],請使用:

  "scaleType": "Default"

針對 TargetUtilization,請使用:

  "scaleType": "TargetUtilization",
  "maxInstances": "int",
  "minInstances": "int",
  "pollingInterval": "string",
  "targetUtilizationPercentage": "int"

屬性值

workspaces/onlineEndpoints/deployments

名稱 描述
類型 資源類型 'Microsoft.MachineLearningServices/workspaces/onlineEndpoints/deployments'
apiVersion 資源 API 版本 '2023-04-01'
NAME 資源名稱

請參閱如何在 JSON ARM 樣本中設定子資源的名稱和類型。
字串 (必要)
location 資源所在的地理位置 字串 (必要)
tags 資源標籤。 標記名稱和值的字典。 請參閱 範本中的標記
sku 自動調整 ARM 合約所需的 SKU 詳細數據。 Sku
kind 入口網站/工具等用來轉譯相同類型資源的不同UX體驗的元數據。 字串
身分識別 受控服務識別 (系統指派和/或使用者指派的身分識別) ManagedServiceIdentity
properties [必要]實體的其他屬性。 OnlineDeploymentProperties (必要)

ManagedServiceIdentity

名稱 描述
類型 (允許 SystemAssigned 和 UserAssigned 類型的受控服務識別類型) 。 'None'
'SystemAssigned'
'SystemAssigned,UserAssigned'
'UserAssigned' (必要)
userAssignedIdentities 與資源相關聯的使用者指派身分識別集。 userAssignedIdentities 字典索引鍵的格式為 ARM 資源標識符:'/subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/providers/Microsoft.ManagedIdentity/userAssignedIdentities/{identityName}。 字典值可以是要求中 () 的空白物件 {} 。 UserAssignedIdentities

UserAssignedIdentities

名稱 描述
{自定義屬性} UserAssignedIdentity

UserAssignedIdentity

此物件不包含部署期間要設定的任何屬性。 所有屬性都是 ReadOnly。

OnlineDeploymentProperties

名稱 描述
appInsightsEnabled 如果為 true,則啟用 Application Insights 記錄。 bool
codeConfiguration 端點部署的程式代碼組態。 CodeConfiguration
description 端點部署的描述。 字串
egressPublicNetworkAccess 如果已啟用,則允許輸出公用網路存取。 如果停用,這會建立安全的輸出。 預設值:已啟用。 'Disabled'
'Enabled'
environmentId 端點部署之環境規格的 ARM 資源識別碼或 AssetId。 字串
environmentVariables 部署的環境變數組態。 EndpointDeploymentPropertiesBaseEnvironmentVariables
instanceType 計算實例類型。 字串
livenessProbe 即時探查會定期監視容器的健康情況。 ProbeSettings
model 模型的 URI 路徑。 字串
modelMountPath 在自定義容器中掛接模型的路徑。 字串
properties 屬性字典。 您可以新增屬性,但無法移除或改變。 EndpointDeploymentPropertiesBaseProperties
readinessProbe 整備探查會驗證容器是否已準備好提供流量。 屬性和預設值與即時探查相同。 ProbeSettings
requestSettings 要求部署的設定。 OnlineRequestSettings
scaleSettings 部署的調整設定。
如果它是 Null 或未提供,則為
默認為 KubernetesOnlineDeployment 的 TargetUtilizationScaleSettings
和 至 ManagedOnlineDeployment 的 DefaultScaleSettings。
OnlineScaleSettings
endpointComputeType 設定物件類型 Kubernetes
受控 (必要)

CodeConfiguration

名稱 描述
codeId 程式代碼資產的 ARM 資源識別碼。 字串
scoringScript [必要]在啟動時執行的腳本。 例如 “score.py” 字串 (必要)

約束:
最小長度 = 1
模式 = [a-zA-Z0-9_]

EndpointDeploymentPropertiesBaseEnvironmentVariables

名稱 描述
{自定義屬性} 字串
{自定義屬性} 字串

ProbeSettings

名稱 描述
failureThreshold 傳回狀況不良狀態之前允許的失敗次數。 int
initialDelay ISO 8601 格式的第一個探查之前的延遲。 字串
Period ISO 8601 格式探查之間的時間長度。 字串
successThreshold 傳回狀況良好的狀態之前,成功探查的數目。 int
timeout ISO 8601 格式的探查逾時。 字串

EndpointDeploymentPropertiesBaseProperties

名稱 描述
{自定義屬性} 字串
{自定義屬性} 字串

OnlineRequestSettings

名稱 描述
maxConcurrentRequestsPerInstance 每個部署允許的每個節點並行要求數目上限。 預設值為 1。 int
maxQueueWait 要求會以 ISO 8601 格式保留在佇列中的最長時間量。
預設為 500 毫秒。
字串
requestTimeout ISO 8601 格式的評分逾時。
預設為 5000 毫秒。
字串

OnlineScaleSettings

名稱 描述
scaleType 設定物件類型 預設值
需要 TargetUtilization ()

DefaultScaleSettings

名稱 描述
scaleType [必要]部署調整演算法的類型 必要) 'Default' (

TargetUtilizationScaleSettings

名稱 描述
scaleType [必要]部署調整演算法的類型 'TargetUtilization' (必要)
maxInstances 部署可調整的實例數目上限。 配額將保留給max_instances。 int
minInstances 要一律存在的實例數目下限。 int
pollingInterval ISO 8691 格式的輪詢間隔。 僅支援有效位數為秒數的持續時間。 字串
targetUtilizationPercentage 自動調整程式的目標 CPU 使用量。 int

KubernetesOnlineDeployment

名稱 描述
endpointComputeType [必要]端點的計算類型。 必要) 'Kubernetes' (
containerResourceRequirements 容器的資源需求 (cpu 和記憶體) 。 ContainerResourceRequirements

ContainerResourceRequirements

名稱 描述
containerResourceLimits 容器資源限制資訊: ContainerResourceSettings
containerResourceRequests 容器資源要求資訊: ContainerResourceSettings

ContainerResourceSettings

名稱 描述
cpu 容器的 vCPU 要求/限制數目。 其他資訊:
https://kubernetes.io/docs/concepts/configuration/manage-compute-resources-container/
字串
gpu 容器的 Nvidia GPU 卡要求/限制數目。 其他資訊:
https://kubernetes.io/docs/concepts/configuration/manage-compute-resources-container/
字串
記憶體 容器的記憶體大小要求/限制。 其他資訊:
https://kubernetes.io/docs/concepts/configuration/manage-compute-resources-container/
字串

ManagedOnlineDeployment

名稱 描述
endpointComputeType [必要]端點的計算類型。 必要) 'Managed' (

SKU

名稱 描述
處理能力 如果 SKU 支援相應放大/縮小,則應該包含容量整數。 如果資源無法相應放大/縮小,可能會省略此專案。 int
family 如果服務有不同層代的硬體,則針對相同的SKU,則可以在這裡擷取。 字串
NAME SKU 的名稱。 例如 - P3。 通常是字母+數字代碼 字串 (必要)
{1}size{2} SKU 大小。 當名稱欄位是階層和其他一些值的組合時,這會是獨立程序代碼。 字串
tier 如果服務有多個層級,但 PUT 上不需要,則資源提供者必須實作此字段。 'Basic'
'Free'
'Premium'
'Standard'

Terraform (AzAPI 提供者) 資源定義

工作區/onlineEndpoints/deployments 資源類型可以使用目標作業來部署:

  • 資源群組

如需每個 API 版本中已變更屬性的清單,請參閱 變更記錄

資源格式

若要建立 Microsoft.MachineLearningServices/workspaces/onlineEndpoints/deployments 資源,請將下列 Terraform 新增至您的範本。

resource "azapi_resource" "symbolicname" {
  type = "Microsoft.MachineLearningServices/workspaces/onlineEndpoints/deployments@2023-04-01"
  name = "string"
  location = "string"
  parent_id = "string"
  tags = {
    tagName1 = "tagValue1"
    tagName2 = "tagValue2"
  }
  identity {
    type = "string"
    identity_ids = []
  }
  body = jsonencode({
    properties = {
      appInsightsEnabled = bool
      codeConfiguration = {
        codeId = "string"
        scoringScript = "string"
      }
      description = "string"
      egressPublicNetworkAccess = "string"
      environmentId = "string"
      environmentVariables = {
        {customized property} = "string"
        {customized property} = "string"
      }
      instanceType = "string"
      livenessProbe = {
        failureThreshold = int
        initialDelay = "string"
        period = "string"
        successThreshold = int
        timeout = "string"
      }
      model = "string"
      modelMountPath = "string"
      properties = {
        {customized property} = "string"
        {customized property} = "string"
      }
      readinessProbe = {
        failureThreshold = int
        initialDelay = "string"
        period = "string"
        successThreshold = int
        timeout = "string"
      }
      requestSettings = {
        maxConcurrentRequestsPerInstance = int
        maxQueueWait = "string"
        requestTimeout = "string"
      }
      scaleSettings = {
        scaleType = "string"
        // For remaining properties, see OnlineScaleSettings objects
      }
      endpointComputeType = "string"
      // For remaining properties, see OnlineDeploymentProperties objects
    }
    sku = {
      capacity = int
      family = "string"
      name = "string"
      size = "string"
      tier = "string"
    }
    kind = "string"
  })
}

OnlineDeploymentProperties 物件

設定 endpointComputeType 屬性以指定對象的類型。

針對 Kubernetes,請使用:

  endpointComputeType = "Kubernetes"
  containerResourceRequirements = {
    containerResourceLimits = {
      cpu = "string"
      gpu = "string"
      memory = "string"
    }
    containerResourceRequests = {
      cpu = "string"
      gpu = "string"
      memory = "string"
    }
  }

針對 Managed,請使用:

  endpointComputeType = "Managed"

OnlineScaleSettings 物件

設定 scaleType 屬性以指定物件的類型。

針對 [預設],請使用:

  scaleType = "Default"

針對 TargetUtilization,請使用:

  scaleType = "TargetUtilization"
  maxInstances = int
  minInstances = int
  pollingInterval = "string"
  targetUtilizationPercentage = int

屬性值

workspaces/onlineEndpoints/deployments

名稱 描述
類型 資源類型 “Microsoft.MachineLearningServices/workspaces/onlineEndpoints/deployments@2023-04-01”
NAME 資源名稱 字串 (必要)
location 資源所在的地理位置 字串 (必要)
parent_id 此資源之父系的資源標識碼。 類型資源的標識碼: onlineEndpoints
tags 資源標籤。 標記名稱和值的字典。
sku 自動調整 ARM 合約所需的 SKU 詳細數據。 Sku
kind 入口網站/工具等用來轉譯相同類型資源的不同UX體驗的元數據。 字串
身分識別 受控服務識別 (系統指派和/或使用者指派的身分識別) ManagedServiceIdentity
properties [必要]實體的其他屬性。 OnlineDeploymentProperties (必要)

ManagedServiceIdentity

名稱 描述
類型 (允許 SystemAssigned 和 UserAssigned 類型的受控服務識別類型) 。 “SystemAssigned”
“SystemAssigned,UserAssigned”
) 所需的 「UserAssigned」 (
identity_ids 與資源相關聯的使用者指派身分識別集。 userAssignedIdentities 字典索引鍵的格式為 ARM 資源標識符:'/subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/providers/Microsoft.ManagedIdentity/userAssignedIdentities/{identityName}。 字典值可以是要求中 () 的空白物件 {} 。 使用者身分識別標識碼的陣列。

UserAssignedIdentities

名稱 描述
{自定義屬性} UserAssignedIdentity

UserAssignedIdentity

此物件不包含部署期間要設定的任何屬性。 所有屬性都是 ReadOnly。

OnlineDeploymentProperties

名稱 描述
appInsightsEnabled 如果為 true,則啟用 Application Insights 記錄。 bool
codeConfiguration 端點部署的程式代碼組態。 CodeConfiguration
description 端點部署的描述。 字串
egressPublicNetworkAccess 如果啟用,則允許輸出公用網路存取。 如果為 Disabled,這會建立安全的輸出。 預設值:已啟用。 “Disabled”
“Enabled”
environmentId 端點部署之環境規格的 ARM 資源識別碼或 AssetId。 字串
environmentVariables 部署的環境變數組態。 EndpointDeploymentPropertiesBaseEnvironmentVariables
instanceType 計算實例類型。 字串
livenessProbe 活躍度探查會定期監視容器的健康情況。 ProbeSettings
model 模型的 URI 路徑。 字串
modelMountPath 在自定義容器中掛接模型的路徑。 字串
properties 屬性字典。 您可以新增屬性,但無法移除或改變。 EndpointDeploymentPropertiesBaseProperties
readinessProbe 整備探查會驗證容器是否已準備好提供流量。 屬性和預設值與活躍度探查相同。 ProbeSettings
requestSettings 要求部署的設定。 OnlineRequestSettings
scaleSettings 部署的調整設定。
如果它是 Null 或未提供,則為
默認為 KubernetesOnlineDeployment 的 TargetUtilizationScaleSettings
和 至 ManagedOnlineDeployment 的 DefaultScaleSettings。
OnlineScaleSettings
endpointComputeType 設定物件類型 Kubernetes
受控 (必要)

CodeConfiguration

名稱 描述
codeId 程式代碼資產的 ARM 資源識別碼。 字串
scoringScript [必要]在啟動時執行的腳本。 例如 “score.py” 需要字串 ()

約束:
最小長度 = 1
模式 = [a-zA-Z0-9_]

EndpointDeploymentPropertiesBaseEnvironmentVariables

名稱 描述
{自定義屬性} 字串
{自定義屬性} 字串

ProbeSettings

名稱 描述
failureThreshold 傳回狀況不良狀態之前允許的失敗次數。 int
initialDelay ISO 8601 格式的第一個探查之前的延遲。 字串
Period ISO 8601 格式探查之間的時間長度。 字串
successThreshold 傳回狀況良好狀態之前成功的探查數目。 int
timeout ISO 8601 格式的探查逾時。 字串

EndpointDeploymentPropertiesBaseProperties

名稱 描述
{自定義屬性} 字串
{自定義屬性} 字串

OnlineRequestSettings

名稱 描述
maxConcurrentRequestsPerInstance 每個部署允許的每個節點並行要求數目上限。 預設值為 1。 int
maxQueueWait 要求會維持在 ISO 8601 格式的佇列中的最大時間量。
預設值為 500 毫秒。
字串
requestTimeout ISO 8601 格式的評分逾時。
預設值為 5000 毫秒。
字串

OnlineScaleSettings

名稱 描述
scaleType 設定物件類型 預設值
TargetUtilization (必要)

DefaultScaleSettings

名稱 描述
scaleType [必要]部署調整演算法的類型 「預設」 (必要)

TargetUtilizationScaleSettings

名稱 描述
scaleType [必要]部署調整演算法的類型 必要) “TargetUtilization” (
maxInstances 部署可調整的實例數目上限。 配額將會保留給max_instances。 int
minInstances 一律存在的實例數目下限。 int
pollingInterval ISO 8691 格式的輪詢間隔。 僅支援有效位數為秒數的持續時間。 字串
targetUtilizationPercentage 自動調整程式的目標 CPU 使用量。 int

KubernetesOnlineDeployment

名稱 描述
endpointComputeType [必要]端點的計算類型。 ) 所需的 「Kubernetes」 (
containerResourceRequirements 容器的資源需求 (cpu 和記憶體) 。 ContainerResourceRequirements

ContainerResourceRequirements

名稱 描述
containerResourceLimits 容器資源限制資訊: ContainerResourceSettings
containerResourceRequests 容器資源要求資訊: ContainerResourceSettings

ContainerResourceSettings

名稱 描述
cpu 容器的 vCPU 要求/限制數目。 其他資訊:
https://kubernetes.io/docs/concepts/configuration/manage-compute-resources-container/
字串
gpu 容器的 Nvidia GPU 卡要求/限制數目。 其他資訊:
https://kubernetes.io/docs/concepts/configuration/manage-compute-resources-container/
字串
記憶體 容器的記憶體大小要求/限制。 其他資訊:
https://kubernetes.io/docs/concepts/configuration/manage-compute-resources-container/
字串

ManagedOnlineDeployment

名稱 描述
endpointComputeType [必要]端點的計算類型。 「受控」 (必要)

SKU

名稱 描述
處理能力 如果 SKU 支援相應放大/縮小,則應該包含容量整數。 如果資源無法相應放大/縮小,則可能會省略此專案。 int
family 如果服務有不同層代的硬體,則針對相同的SKU,則可以在這裡擷取。 字串
NAME SKU 的名稱。 例如 - P3。 通常是字母+數字代碼 字串 (必要)
{1}size{2} SKU 大小。 當名稱欄位是階層和其他一些值的組合時,這會是獨立程序代碼。 字串
tier 如果服務有多個層級,但 PUT 上不需要,則資源提供者必須實作此字段。 "Basic"
“Free”
“Premium”
“Standard”