managed_compute_deployments
Creates, updates, deletes, gets or lists a managed_compute_deployments resource.
Overview
| Name | managed_compute_deployments |
| Type | Resource |
| Id | azure.cognitive_services.managed_compute_deployments |
Fields
The following fields are returned by SELECT queries:
- get
- list
| Name | Datatype | Description |
|---|---|---|
id | string | Fully qualified resource ID for the resource. Ex - /subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/providers/{resourceProviderNamespace}/{resourceType}/{resourceName}. |
name | string | The name of the resource. |
acceleratorType | string | Accelerator type (e.g., H100_80GB). Optional on creation; immutable after creation. |
acceleratorsPerInstance | integer | Read-only. Number of accelerators (GPUs) consumed by each model instance, sourced from the deployment template. |
capabilities | object | Deployment capabilities represented as key-value pairs. Example: { assetsV2: "true" }. |
computeId | string | Foundry compute ARM resource ID for VM-backed managed compute deployments. Required when sku.name is VmManagedCompute; immutable after creation. |
deploymentTemplate | string | Deployment template identifier. Optional on creation. Accepts an AzureML Registry deployment template URI or a project-scoped deployment template path for VmManagedCompute. Examples: azureml://registries/{registry}/deploymenttemplates/{template}/versions/{version}, projects/{project}/deploymentTemplates/{template}/versions/{version}. |
etag | string | Resource Etag. |
model | string | AzureML Registry model asset URI. Required on creation; immutable after creation. Example: azureml://registries/{registry}/models/{model}/versions/{version}. Required. |
priority | string | Scheduling priority for VM-backed managed compute deployments. Immutable after creation. |
provisioningDetails | object | Read-only. Status message and timestamp from the last provisioning operation. |
provisioningState | string | Read-only. Current provisioning state. Known values are: "Accepted", "Creating", "Deleting", "Moving", "Failed", "Succeeded", "Canceled", and "ResolvingDNS". (Accepted, Creating, Deleting, Moving, Failed, Succeeded, Canceled, ResolvingDNS) |
routes | object | Read-only. Inference route paths relative to the account endpoint. Populated when provisioningState is Succeeded. |
sku | object | The resource model definition representing SKU. |
systemData | object | Azure Resource Manager metadata containing createdBy and modifiedBy information. |
totalAccelerators | integer | Read-only. Total accelerators allocated: sku.capacity (instances) x acceleratorsPerInstance. |
type | string | The type of the resource. E.g. "Microsoft.Compute/virtualMachines" or "Microsoft.Storage/storageAccounts". |
versionUpgradeOption | string | Template auto-upgrade policy. Defaults to OnceNewDefaultVersionAvailable. Known values are: "OnceNewDefaultVersionAvailable", "OnceCurrentVersionExpired", and "NoAutoUpgrade". (OnceNewDefaultVersionAvailable, OnceCurrentVersionExpired, NoAutoUpgrade) |
| Name | Datatype | Description |
|---|---|---|
id | string | Fully qualified resource ID for the resource. Ex - /subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/providers/{resourceProviderNamespace}/{resourceType}/{resourceName}. |
name | string | The name of the resource. |
acceleratorType | string | Accelerator type (e.g., H100_80GB). Optional on creation; immutable after creation. |
acceleratorsPerInstance | integer | Read-only. Number of accelerators (GPUs) consumed by each model instance, sourced from the deployment template. |
capabilities | object | Deployment capabilities represented as key-value pairs. Example: { assetsV2: "true" }. |
computeId | string | Foundry compute ARM resource ID for VM-backed managed compute deployments. Required when sku.name is VmManagedCompute; immutable after creation. |
deploymentTemplate | string | Deployment template identifier. Optional on creation. Accepts an AzureML Registry deployment template URI or a project-scoped deployment template path for VmManagedCompute. Examples: azureml://registries/{registry}/deploymenttemplates/{template}/versions/{version}, projects/{project}/deploymentTemplates/{template}/versions/{version}. |
etag | string | Resource Etag. |
model | string | AzureML Registry model asset URI. Required on creation; immutable after creation. Example: azureml://registries/{registry}/models/{model}/versions/{version}. Required. |
priority | string | Scheduling priority for VM-backed managed compute deployments. Immutable after creation. |
provisioningDetails | object | Read-only. Status message and timestamp from the last provisioning operation. |
provisioningState | string | Read-only. Current provisioning state. Known values are: "Accepted", "Creating", "Deleting", "Moving", "Failed", "Succeeded", "Canceled", and "ResolvingDNS". (Accepted, Creating, Deleting, Moving, Failed, Succeeded, Canceled, ResolvingDNS) |
routes | object | Read-only. Inference route paths relative to the account endpoint. Populated when provisioningState is Succeeded. |
sku | object | The resource model definition representing SKU. |
systemData | object | Azure Resource Manager metadata containing createdBy and modifiedBy information. |
totalAccelerators | integer | Read-only. Total accelerators allocated: sku.capacity (instances) x acceleratorsPerInstance. |
type | string | The type of the resource. E.g. "Microsoft.Compute/virtualMachines" or "Microsoft.Storage/storageAccounts". |
versionUpgradeOption | string | Template auto-upgrade policy. Defaults to OnceNewDefaultVersionAvailable. Known values are: "OnceNewDefaultVersionAvailable", "OnceCurrentVersionExpired", and "NoAutoUpgrade". (OnceNewDefaultVersionAvailable, OnceCurrentVersionExpired, NoAutoUpgrade) |
Methods
The following methods are available for this resource:
| Name | Accessible by | Required Params | Optional Params | Description |
|---|---|---|---|---|
get | select | resource_group_name, account_name, deployment_name, subscription_id | Gets the specified managed compute deployment associated with the Cognitive Services account. | |
list | select | resource_group_name, account_name, subscription_id | Gets the managed compute deployments associated with the Cognitive Services account. | |
create_or_update | insert | resource_group_name, account_name, deployment_name, subscription_id | Creates or updates a managed compute deployment associated with the Cognitive Services account. | |
update | update | resource_group_name, account_name, deployment_name, subscription_id | Updates the specified managed compute deployment associated with the Cognitive Services account. | |
create_or_update | replace | resource_group_name, account_name, deployment_name, subscription_id | Creates or updates a managed compute deployment associated with the Cognitive Services account. | |
delete | delete | resource_group_name, account_name, deployment_name, subscription_id | Deletes the specified managed compute deployment associated with the Cognitive Services account. |
Parameters
Parameters can be passed in the WHERE clause of a query. Check the Methods section to see which parameters are required or optional for each operation.
| Name | Datatype | Description |
|---|---|---|
account_name | string | The name of Cognitive Services account. Required. |
deployment_name | string | The name of the managed compute deployment associated with the Cognitive Services Account. Required. |
resource_group_name | string | The name of the resource group. The name is case insensitive. Required. |
subscription_id | string |
SELECT examples
- get
- list
Gets the specified managed compute deployment associated with the Cognitive Services account.
SELECT
id,
name,
acceleratorType,
acceleratorsPerInstance,
capabilities,
computeId,
deploymentTemplate,
etag,
model,
priority,
provisioningDetails,
provisioningState,
routes,
sku,
systemData,
totalAccelerators,
type,
versionUpgradeOption
FROM azure.cognitive_services.managed_compute_deployments
WHERE resource_group_name = '{{ resource_group_name }}' -- required
AND account_name = '{{ account_name }}' -- required
AND deployment_name = '{{ deployment_name }}' -- required
AND subscription_id = '{{ subscription_id }}' -- required
;
Gets the managed compute deployments associated with the Cognitive Services account.
SELECT
id,
name,
acceleratorType,
acceleratorsPerInstance,
capabilities,
computeId,
deploymentTemplate,
etag,
model,
priority,
provisioningDetails,
provisioningState,
routes,
sku,
systemData,
totalAccelerators,
type,
versionUpgradeOption
FROM azure.cognitive_services.managed_compute_deployments
WHERE resource_group_name = '{{ resource_group_name }}' -- required
AND account_name = '{{ account_name }}' -- required
AND subscription_id = '{{ subscription_id }}' -- required
;
INSERT examples
- create_or_update
- Manifest
Creates or updates a managed compute deployment associated with the Cognitive Services account.
INSERT INTO azure.cognitive_services.managed_compute_deployments (
properties,
sku,
resource_group_name,
account_name,
deployment_name,
subscription_id
)
SELECT
'{{ properties }}',
'{{ sku }}',
'{{ resource_group_name }}',
'{{ account_name }}',
'{{ deployment_name }}',
'{{ subscription_id }}'
RETURNING
id,
name,
etag,
properties,
sku,
systemData,
type
;
# Description fields are for documentation purposes
- name: managed_compute_deployments
props:
- name: resource_group_name
value: "{{ resource_group_name }}"
description: Required parameter for the managed_compute_deployments resource.
- name: account_name
value: "{{ account_name }}"
description: Required parameter for the managed_compute_deployments resource.
- name: deployment_name
value: "{{ deployment_name }}"
description: Required parameter for the managed_compute_deployments resource.
- name: subscription_id
value: "{{ subscription_id }}"
description: Required parameter for the managed_compute_deployments resource.
- name: properties
description: |
Properties of the Cognitive Services managed compute deployment.
value:
model: "{{ model }}"
deploymentTemplate: "{{ deploymentTemplate }}"
acceleratorType: "{{ acceleratorType }}"
versionUpgradeOption: "{{ versionUpgradeOption }}"
capabilities: "{{ capabilities }}"
computeId: "{{ computeId }}"
priority: "{{ priority }}"
acceleratorsPerInstance: {{ acceleratorsPerInstance }}
totalAccelerators: {{ totalAccelerators }}
provisioningState: "{{ provisioningState }}"
provisioningDetails:
message: "{{ message }}"
lastOperationTimestamp: "{{ lastOperationTimestamp }}"
routes:
chatCompletionsScoringPath: "{{ chatCompletionsScoringPath }}"
swagger: "{{ swagger }}"
messagesApiScoringPath: "{{ messagesApiScoringPath }}"
- name: sku
description: |
The resource model definition representing SKU.
value:
name: "{{ name }}"
tier: "{{ tier }}"
size: "{{ size }}"
family: "{{ family }}"
capacity: {{ capacity }}
UPDATE examples
- update
Updates the specified managed compute deployment associated with the Cognitive Services account.
UPDATE azure.cognitive_services.managed_compute_deployments
SET
sku = '{{ sku }}'
WHERE
resource_group_name = '{{ resource_group_name }}' --required
AND account_name = '{{ account_name }}' --required
AND deployment_name = '{{ deployment_name }}' --required
AND subscription_id = '{{ subscription_id }}' --required
RETURNING
id,
name,
etag,
properties,
sku,
systemData,
type;
REPLACE examples
- create_or_update
Creates or updates a managed compute deployment associated with the Cognitive Services account.
REPLACE azure.cognitive_services.managed_compute_deployments
SET
properties = '{{ properties }}',
sku = '{{ sku }}'
WHERE
resource_group_name = '{{ resource_group_name }}' --required
AND account_name = '{{ account_name }}' --required
AND deployment_name = '{{ deployment_name }}' --required
AND subscription_id = '{{ subscription_id }}' --required
RETURNING
id,
name,
etag,
properties,
sku,
systemData,
type;
DELETE examples
- delete
Deletes the specified managed compute deployment associated with the Cognitive Services account.
DELETE FROM azure.cognitive_services.managed_compute_deployments
WHERE resource_group_name = '{{ resource_group_name }}' --required
AND account_name = '{{ account_name }}' --required
AND deployment_name = '{{ deployment_name }}' --required
AND subscription_id = '{{ subscription_id }}' --required
;