Based on your existing cloud quotas, your default compute size has been reduced. View or increase your quotas in Azure portal.

Hi @sruthi salavadi

Based on your existing cloud quotas, your default compute size has been reduced. View or increase your quotas in Azure portal." The reply on the thread pointed you to a related discussion, so I want to give you the complete path here rather than send you hopping between threads including the one check that determines whether a quota request will work for you at all.

Root cause likely, pending two data points

This message is Azure Databricks reporting that the compute size you selected exceeds the vCPU quota available to your subscription in your workspace's region. Databricks provisions standard Azure VMs and inherits Azure's quota model, which is enforced per subscription, per region, and per VM family. Rather than fail the cluster, the UI reduces the default node size to fit and raises this banner.

there are two distinct causes that produce this identical message, and your thread doesn't yet contain the information to separate them. Either your subscription is a Free Trial, Student, or Pass offer which are not eligible for quota increases, and can report an approved vCPU quota of 0 even at zero usage or you're on a paid offer and the node SKU simply exceeds the family limit. I've ruled out a service defect: no service health advisory or known-issue record matched this string, and Microsoft's published documentation states plainly that quotas may be adjusted downward where allocated quota isn't utilized. That documented behavior is the most probable reason your size dropped without any change on your side.

Resolution steps:

Step 1: Identify your subscription offer type. Azure Portal → Subscriptions → your subscription → Overview → Offer. Why this matters Free Trial, Azure for Students, and Azure Pass subscriptions cannot receive quota increases, so Steps 3 and 4 will be rejected. This single lookup determines your entire path. How to verify: the Offer field names your plan explicitly.

Step 2 (only if Step 1 shows a trial/student/pass offer): Convert to Pay-As-You-Go and remove the spending limit. Why this matters: Azure Databricks' documented prerequisite is a subscription that isn't a Free Trial; conversion is the only supported route to usable compute. How to verify: the Overview blade shows Pay-As-You-Go with no spending limit, and Microsoft.ManagedIdentity reads Registered under Subscriptions → Resource providers. You'll need Contributor or Owner rights to complete this.

Step 3: Measure the actual gap. Record your cluster's Driver and Worker node types, then run az vm list-usage --location "<your region>" -o table (or Get-AzVMUsage -Location "<your region>" in PowerShell). Why this matters: Azure checks both the VM-family limit and Total Regional vCPUs, and it counts allocated and deallocated cores alike stopped VMs in that region are still consuming your quota. Sizing the request against your autoscale maximum avoids a second round-trip. How to verify: CurrentValue plus your cluster's required cores must be at or below Limit on both rows.

Step 4: Submit the quota increase. Fastest route is Azure Home → My quotas, where you can submit an amount or usage percentage directly. Alternatively: Help + support → New support request → Issue type Service and subscription limits (quotas) → Quota type Compute-VM (cores-vCPUs) → select VM family and region. Why this matters: this is the only supported way to lift an adjustable quota, and there is no cost to requesting one. How to verify: the Limit column on Usage + quotas reflects the new value; restart the compute and confirm the banner is gone.

If you have further questions regarding this answer, feel free to click "Comment". If you find the answer helpful, please click "upvote". This helps the community by allowing others with similar queries to easily find the solution.

Refer to https://learn.microsoft.com/en-us/answers/questions/5551265/azure-subscription-quota-limitation-for-databricks

That warning means your Databricks workspace tried to start a cluster/compute that requires more vCPUs/GPUs than your subscription quota allows in the target region. Azure enforces quota limits per subscription, per region, per VM family (Standard_DSv3, Standard_NCv2, etc.), and Databricks inherits those limits.

Steps to fix:

1. Check which VM family is being used in the Databricks workspace
2. Check quotas in the Azure portal
   1. Go to Azure Portal → Subscriptions → [Your Subscription] → Usage + quotas.
   2. Filter by the region where your Databricks workspace is deployed.
   3. Look for the VM family quota (e.g., Standard DSv3 Family vCPUs).
3. Request a quota increase
   • In the Usage + quotas blade, select the VM family → Request increase.
   • Submit a support request to raise the vCPU limit to the number needed for your Databricks cluster (factor in autoscaling, workers × cores).
   • Microsoft usually reviews and approves increases within a few hours to a couple of days.

If the above response helps answer your question, remember to "Accept Answer" so that others in the community facing similar issues can easily find the solution. Your contribution is highly appreciated.

hth

Marcin