Manage and request quotas for Azure resources

As with other Azure services, there are limits on certain resources associated with the Azure Machine Learning service. These limits range from a cap on the number of workspaces you can create to limits on the actual underlying compute that gets used for model training or inference/scoring.

This article gives more details on the pre-configured limits on various Azure resources for your subscription and also contains handy links to request quota enhancements for each type of resource. These limits are put in place to prevent budget over-runs due to fraud, and to honor Azure capacity constraints.

Keep these quotas in mind as you design and scale up your Azure Machine Learning service resources for production workloads. For example, if your cluster doesn't reach the target number of nodes you specified, then you might have reached an Azure Machine Learning Compute cores limit for your subscription. If you want to raise the limit or quota above the Default Limit, open an online customer support request at no charge. The limits can't be raised above the Maximum Limit value shown in the following tables due to Azure Capacity constraints. If there is no Maximum Limit column, then the resource doesn't have adjustable limits.

Special considerations

• A quota is a credit limit, not a capacity guarantee. If you have large-scale capacity needs, contact Azure support.
• Your quota is shared across all the services in your subscriptions including Azure Machine Learning service. The only exception is Azure Machine Learning compute which has a separate quota from the core compute quota. Be sure to calculate the quota usage across all services when evaluating your capacity needs.
• Default limits vary by offer Category Type, such as Free Trial, Pay-As-You-Go, and series, such as Dv2, F, G, and so on.

Default resource quotas

Here is a breakdown of the quota limits by various resource types within your Azure subscription.

 Important

Limits are subject to change. The latest can always be found at the service-level quota document for all of Azure.

Virtual machines

There is a limit on the number of virtual machines you can provision on an Azure subscription across your services or in a standalone manner. This limit is at the region level both on the total cores and also on a per family basis.

It is important to emphasize that virtual machine cores have a regional total limit and a regional per size series (Dv2, F, etc.) limit that are separately enforced. For example, consider a subscription with a US East total VM core limit of 30, an A series core limit of 30, and a D series core limit of 30. This subscription would be allowed to deploy 30 A1 VMs, or 30 D1 VMs, or a combination of the two not to exceed a total of 30 cores (for example, 10 A1 VMs and 20 D1 VMs).

Resource	Default limit	Maximum limit
VMs per subscription	25,0001 per region.	25,000 per region.
VM total cores per subscription	201 per region.	Contact support.
VM per series, such as Dv2 and F, cores per subscription	201 per region.	Contact support.
Coadministrators per subscription	Unlimited.	Unlimited.
Storage accounts per region per subscription	250	250
Resource groups per subscription	980	980
Availability sets per subscription	2,000 per region.	2,000 per region.
Azure Resource Manager API request size	4,194,304 bytes.	4,194,304 bytes.
Tags per subscription2	Unlimited.	Unlimited.
Unique tag calculations per subscription2	10,000	10,000
Cloud services per subscription	N/A3	N/A3
Affinity groups per subscription	N/A3	N/A3
Subscription-level deployments per location	8004	800

1Default limits vary by offer category type, such as Free Trial and Pay-As-You-Go, and by series, such as Dv2, F, and G. For example, the default for Enterprise Agreement subscriptions is 350.

2You can apply an unlimited number of tags per subscription. The number of tags per resource or resource group is limited to 50. Resource Manager returns a list of unique tag name and values in the subscription only when the number of tags is 10,000 or less. You still can find a resource by tag when the number exceeds 10,000.

3These features are no longer required with Azure resource groups and Resource Manager.

4If you reach the limit of 800 deployments, delete deployments from the history that are no longer needed. To delete subscription level deployments, use Remove-AzDeployment or az deployment delete.

 Note

Virtual machine cores have a regional total limit. They also have a limit for regional per-size series, such as Dv2 and F. These limits are separately enforced. For example, consider a subscription with a US East total VM core limit of 30, an A series core limit of 30, and a D series core limit of 30. This subscription can deploy 30 A1 VMs, or 30 D1 VMs, or a combination of the two not to exceed a total of 30 cores. An example of a combination is 10 A1 VMs and 20 D1 VMs.

For a more detailed and up-to-date list of quota limits, check the Azure-wide quota article here.

Azure Machine Learning Compute

For Azure Machine Learning Compute, there is a default quota limit on both the number of cores and number of unique compute resources allowed per region in a subscription. This quota is separate from the VM core quota above and the core limits are not shared currently between the two resource types.

Available resources:

• Dedicated cores per region have a default limit of 24 - 300 depending on your subscription offer type. The number of dedicated cores per subscription can be increased. Contact Azure support to discuss increase options.
• Low-priority cores per region have a default limit of 24 - 300 depending on your subscription offer type. The number of low-priority cores per subscription can be increased. Contact Azure support to discuss increase options.
• Clusters per region have a default limit of 100 and a maximum limit of 200. Contact Azure support if you want to request an increase beyond this limit.
• There are other strict limits which cannot be exceeded once hit.

Resource	Maximum limit
Maximum workspaces per resource group	800
Maximum nodes in a single Azure Machine Learning Compute (AmlCompute) resource	100 nodes
Maximum GPU MPI processes per node	1-4
Maximum GPU workers per node	1-4
Maximum job lifetime	90 days1
Maximum job lifetime on a Low Priority Node	1 day2
Maximum parameter servers per node	1

1 The maximum lifetime refers to the time that a run start and when it finishes. Completed runs persist indefinitely; data for runs not completed within the maximum lifetime is not accessible. 2 Jobs on a Low Priority node could be pre-empted any time there is a capacity constraint. It is recommended to implement checkpointing in your job.

Azure Machine Learning Pipelines

For Azure Machine Learning Pipelines, there is a quota limit on the number of steps in a pipeline and on the number of schedule-based runs of published pipelines per region in a subscription.

• Maximum number of steps allowed in a pipeline is 30,000
• Maximum number of the sum of schedule-based runs and blob pulls for blog-triggered schedules of published pipelines per subscription per month is 100,000

 Note

If you want to increase this limit, contact Microsoft Support.

Container instances

There is also a limit on the number of container instances that you can spin up in a given time period (scoped hourly) or across your entire subscription.

Resource	Default limit
Container groups per subscription	1001
Number of containers per container group	60
Number of volumes per container group	20
Ports per IP	5
Container instance log size - running instance	4 MB
Container instance log size - stopped instance	16 KB or 1,000 lines
Container creates per hour	3001
Container creates per 5 minutes	1001
Container deletes per hour	3001
Container deletes per 5 minutes	1001

1To request a limit increase, create an Azure Support request.

For a more detailed and up-to-date list of quota limits, check the Azure-wide quota article here.

Storage

There is a limit on the number of storage accounts per region as well in a given subscription. The default limit is 200 and includes both Standard and Premium Storage accounts. If you require more than 200 storage accounts in a given region, make a request through Azure Support. The Azure Storage team will review your business case and may approve up to 250 storage accounts for a given region.

Find your quotas

Viewing your quota for various resources, such as Virtual Machines, Storage, Network, is easy through the Azure portal.

1. On the left pane, select All services and then select Subscriptions under the General category.
2. From the list of subscriptions, select the subscription whose quota you are looking for.
   There is a caveat, specifically for viewing the Azure Machine Learning Compute quota. As mentioned above, that quota is separate from the compute quota on your subscription.
3. On the left pane, select Machine Learning service and then select any workspace from the list shown
4. On the next blade, under the Support + troubleshooting section select Usage + quotas to view your current quota limits and usage.
5. Select a subscription to view the quota limits. Remember to filter to the region you are interested in.

Request quota increases

If you want to raise the limit or quota above the default limit, open an online customer support request at no charge.

The limits can't be raised above the maximum limit value shown in the tables. If there is no maximum limit, then the resource doesn't have adjustable limits. This article covers the quota increase process in more detail.

When requesting a quota increase, you need to select the service you are requesting to raise the quota against, which could be services such as Machine Learning service quota, Container instances or Storage quota. In addition for Azure Machine Learning Compute, you can simply click on the Request Quota button while viewing the quota following the steps above.

 Note

Free Trial subscriptions are not eligible for limit or quota increases. If you have a Free Trial subscription, you can upgrade to a Pay-As-You-Go subscription. For more information, see Upgrade Azure Free Trial to Pay-As-You-Go and Free Trial subscription FAQ.