Manage and request quotas for Azure resources
APPLIES TO: Basic edition Enterprise edition (Upgrade to Enterprise edition)
This article provides details on preconfigured limits on Azure resources for your subscription. Also included are instructions on how 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.
As with other Azure services, there are limits on certain resources associated with Azure Machine Learning. These limits range from a cap on the number of workspaces to limits on the actual underlying compute that gets used for model training or inference/scoring.
As you design and scale your Azure Machine Learning resources for production workloads, consider these limits. For example, if your cluster doesn't reach the target number of nodes, then you may 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.
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. 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 VM 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.
Limits are subject to change. The latest can always be found at the service-level quota document for all of Azure.
For each Azure subscription, there is a limit on the number of virtual machines you can have across your services or standalone. This limit is at the region level both on the total cores and also on a per family basis.
Virtual machine cores have a regional total limit and a regional per size series (Dv2, F, etc.) limit, both of which 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.|
|Azure Spot 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.
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 between the two resource types since AmlCompute is a managed service that deploys resources in a hosted-on-behalf-of model.
Dedicated cores per region have a default limit of 24 - 300 depending on your subscription offer type with higher defaults for EA and CSP offer types. The number of dedicated cores per subscription can be increased and is different for each VM family. Certain specialized VM families like NCv2, NCv3, or ND series start with a default of zero cores. Contact Azure support by raising a quota request to discuss increase options.
Low-priority cores per region have a default limit of 100 - 3000 depending on your subscription offer type with higher defaults for EA and CSP offer types. The number of low-priority cores per subscription can be increased and is a single value across VM families. Contact Azure support to discuss increase options.
Clusters per region have a default limit of 200. These are shared between a training cluster and a compute instance (which is considered as a single node cluster for quota purposes). Contact Azure support if you want to request an increase beyond this limit.
There are **other strict limits that cannot be exceeded once hit.
|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||7 days2|
|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 preempted anytime there is a capacity constraint. We recommend you 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
If you want to increase this limit, contact Microsoft Support.
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.
|Container groups per region 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.
There is a limit on the number of storage accounts per region as well in a given subscription. The default limit is 250 and includes both Standard and Premium Storage accounts. If you require more than 250 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.
Workspace level quota
To better manage resource allocations for Amlcompute between various workspaces, we have introduced a feature that allows you to distribute subscription level quotas (by VM family) and configure them at the workspace level. The default behavior is that all workspaces have the same quota as the subscription level quota for any VM family. However, as the number of workspaces increases, and workloads of varying priority start sharing the same resources, users want a way to better share capacity and avoid resource contention issues. Azure Machine Learning provides a solution with its managed compute offering by allowing users to set a maximum quota for a particular VM family on each workspace. This is analogous to distributing your capacity between workspaces, and the users can choose to also over-allocate to drive maximum utilization.
To set quotas at the workspace level, go to any workspace in your subscription, and click on Usages + quotas in the left pane. Then select the Configure quotas tab to view the quotas, expand any VM family, and set a quota limit on any workspace listed under that VM family. Remember that you cannot set a negative value or a value higher than the subscription level quota. Also, as you would observe, by default all workspaces are assigned the entire subscription quota to allow for full utilization of the allocated quota.
This is an Enterprise edition feature only. If you have both a Basic and an Enterprise edition workspace in your subscription, you can use this to only set quotas on your Enterprise workspaces. Your Basic workspaces will continue to have the subscription level quota which is the default behavior.
You need subscription level permissions to set quota at the workspace level. This is enforced so that individual workspace owners do not edit or increase their quotas and start encroaching onto resources set aside for another workspace. Thus a subscription admin is best suited to allocate and distribute these quotas across workspaces.
View your usage and quotas
Viewing your quota for various resources, such as Virtual Machines, Storage, Network, is easy through the Azure portal.
On the left pane, select All services and then select Subscriptions under the General category.
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.
On the left pane, select Machine Learning service and then select any workspace from the list shown
On the next blade, under the Support + troubleshooting section select Usage + quotas to view your current quota limits and usage.
Select a subscription to view the quota limits. Remember to filter to the region you are interested in.
You can now toggle between a subscription level view and a workspace level view:
Subscription view: This allows you to view your usage of core quota by VM family, expanding it by workspace, and further expanding it by the actual cluster names. This view is optimal for quickly getting into the details of core usage for a particular VM family to see the break-up by workspaces and further by the underlying clusters for each of those workspaces. The general convention in this view is (usage/quota), where the usage is the current number of scaled up cores, and quota is the logical maximum number of cores that the resource can scale to. For each workspace, the quota would be the workspace level quota (as explained above) which denotes the maximum number of cores that you can scale to for a particular VM family. For a cluster similarly, the quota is actually the cores corresponding to the maximum number of nodes that the cluster can scale to defined by the max_nodes property.
Workspace view: This allows you to view your usage of core quota by Workspace, expanding it by VM family, and further expanding it by the actual cluster names. This view is optimal for quickly getting into the details of core usage for a particular workspace to see the break-up by VM families and further by the underlying clusters for each of those families.
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 click on the Request Quota button while viewing the quota following the steps above.