Unable to create any compute cluster in Azure Databricks — quota exceeded on all attempts, including minimal/single-node configurations, across multiple regions.

KAUSAL ASHOK MELEDATH 0 Reputation points
2026-09-11T09:00:34.0233333+00:00

Issue: Unable to create any compute cluster in Azure Databricks — quota exceeded on all attempts, including minimal/single-node configurations, across multiple regions.

Details:

I am unable to create a Databricks compute cluster regardless of configuration. I have tried:

A minimal single-node cluster setup

The default "Personal Compute" policy

A custom-created cluster policy

Multiple regions, including a US region and Germany West Central

All attempts fail with the same quota error. Specifically, when selecting node type Standard_D4ds_v5, the portal shows: "This account may not have enough CPU cores to satisfy this request — Estimated available: 0, requested: 4." A previous attempt with node type Standard_DC4as_v5 also failed to start, with error code AZURE_QUOTA_EXCEEDED_EXCEPTION.

Since this fails consistently across different regions, VM series, and cluster policies, this appears to be a subscription-level vCPU quota issue rather than something specific to one region or VM type.

Subscription type: Pay-As-You-Go

Request: Please advise on increasing vCPU quota for my subscription across the regions above, or confirm why a Pay-As-You-Go subscription is showing 0 available vCPUs by default, so I can create a working Databricks cluster.

Regions attempted: [ East US, Germany West Central]

Azure Databricks
Azure Databricks

An Apache Spark-based analytics platform optimized for Azure.

0 comments

2 answers

Sort by: Oldest
Most helpful Newest Oldest

  1. AI answer

    2026-09-11T09:00:56.3033333+00:00

    AZURE_QUOTA_EXCEEDED_EXCEPTION means the cluster launch would exceed the Azure subscription quota for the requested VM family. The Databricks guidance is to check quota usage, identify the exceeded VM family, review running VMs and clusters, and then either request a quota increase, terminate unused resources, or switch to a VM type with available quota.

    Because the portal shows Estimated available: 0, requested: 4 for Standard_D4ds_v5, the immediate checks are:

    1. In Azure portal, go to Subscriptions > Usage + quotas.
    2. Check both of these quota tiers for each attempted region:
      • Total Regional vCPUs
      • The specific VM family quota for the node type, such as the D-series family.
    3. Confirm whether either quota is already fully used. Azure requires both the regional quota and the VM family quota to have available capacity.
    4. Review any running or deallocated VMs and clusters in the subscription. Quota is calculated from cores in use by both allocated and deallocated VMs.

    To verify from CLI, run:

    az vm list-usage --location "East US" -o table
az vm list-usage --location "Germany West Central" -o table

    If quota is exhausted, the supported fixes are:

    1. Request a quota increase for the affected region and VM family.
      • In Azure portal, open Quotas.
      • Select Microsoft.Compute.
      • Filter by the target Location.
      • Find the out-of-quota record.
      • Select the pencil icon.
      • In Request quota increase, enter the New limit and submit.
    2. If the issue is the Total Regional vCPU quota, request a regional vCPU quota increase or use another region.
    3. If there are unused VMs or clusters, delete them to free quota.
    4. If one VM family is exhausted, choose a different VM type that has available quota.

    A Pay-As-You-Go subscription does not guarantee that every region or VM family has available quota by default. Quota is enforced per subscription and per region, and Azure checks both quota and actual capacity separately. Even after quota is increased, deployment can still fail if the selected VM size has no capacity in that region. In that case, try a different VM size, zone, or region.

    For the regions listed, the required action is to inspect the quota records for East US and Germany West Central and submit quota increase requests for the exhausted Microsoft.Compute regional or VM-family quotas. If the Databricks error message includes the minimum new limit required, request at least that amount.

    References:

    AI-generated content may be incorrect. Read our transparency notes for more information.

    Was this answer helpful?

  2. Vinodh247-1375 44,556 Reputation points Volunteer Moderator
    2026-09-11T09:33:30.05+00:00

    The error message certainly points toward a quota-related restriction, but one detail stands out: you are seeing 0 available vCPUs across multiple VM families and multiple regions. That is not the most common pattern for a simple exhausted quota.

    In addition to reviewing the regional and VM-family quotas already mentioned, I would verify whether the subscription has actually been granted compute quota in those regions.

    A useful validation is to check Subscriptions -> Usage + quotas -> Compute and compare:

    • Total Regional vCPUs
    • The specific VM-family quota (Dv5, DCv5, etc.)
    • Current usage versus limit

    If the limit itself is 0 (not just available = 0), that typically indicates the subscription does not currently have quota allocated for that region or VM family rather than quota being consumed by existing resources.

    Since you tested both Standard_D4ds_v5 and Standard_DC4as_v5 and saw similar behaviour across East US and Germany West Central, I would also check whether the same subscription can deploy a small standalone Azure VM in those regions. If a basic VM deployment fails with a similar quota restriction, that helps confirm the problem exists at the Azure Compute subscription layer rather than within Azure Databricks itself.

    Another point worth checking is whether there are any subscription-level spending limits, offer-type restrictions, or recently created subscription constraints that may result in very low or zero initial compute quotas. A PAYG subscription does not automatically guarantee pre-allocated vCPU capacity in every region and VM family.

    Based on the symptoms described, this appears more likely to be a subscription-level compute quota allocation issue than a Databricks cluster policy, cluster configuration, or Spark related problem.

    Help make this community better for everyone: if this answer resolved your issue, please accept it or leave an upvote. If not, share more details in a comment so we can continue the discussion and find the right solution.

    Was this answer helpful?

Your answer

Answers can be marked as 'Accepted' by the question author and 'Recommended' by moderators, which helps users know the answer solved the author's problem.