Azure Databricks Claude Foundation Model API returns Databricks-set rate limit of 0

Question

Azure Databricks Claude Foundation Model API returns Databricks-set rate limit of 0

Winston Martinez 0

I created an Azure Databricks Premium Serverless workspace in East US.

Claude Foundation Model API requests fail with:

PERMISSION_DENIED: The endpoint is temporarily disabled due to a Databricks-set rate limit of 0.

Workspace: mypersoulases

The workspace is Premium and Serverless. Claude models are visible, but invocation is blocked.

Please check whether this workspace is on a restricted/trial trust tier and whether it needs to be moved to a payable/fully enabled tier for Databricks-hosted Anthropic Claude models.

Please confirm the current Databricks-managed rate limit/entitlement for Claude and whether it can be enabled for this workspace.

Charan Venkat Sai Thambi 0 Reputation points Microsoft External Staff Moderator

2026-09-02T10:43:21.0633333+00:00
Hi @Winston Martinez
Thank you for the details.

The error message:

PERMISSION_DENIED: The endpoint is temporarily disabled due to a Databricks-set rate limit of 0.

indicates that access to the selected foundation model endpoint is currently blocked by a Databricks-managed quota or entitlement setting rather than by the Azure Databricks workspace SKU itself.

A few points to verify:

Confirm that the workspace is attached to a paid Azure subscription and that billing is fully enabled.

Verify that AI/Model Serving features and Foundation Models are supported and enabled for your workspace region.

Review the Model Serving and Foundation Model entitlement settings in your Databricks account console, if available.

Check whether other foundation models can be invoked successfully from the same workspace, or whether the issue is specific to Claude-hosted endpoints.

Because the reported rate limit is explicitly being returned by Databricks, Microsoft Learn support and community moderators do not have visibility into the Databricks-managed trust tier, entitlement status, or model-specific quotas assigned to an individual workspace. We therefore cannot confirm whether the workspace is on a restricted trust tier or manually enable Claude model quotas.

2 answers

Your answer

Charan Venkat Sai Thambi 0 Reputation points Microsoft External Staff Moderator

2026-09-02T10:43:21.0633333+00:00

Hi @Winston Martinez
Thank you for the details.

The error message:

PERMISSION_DENIED: The endpoint is temporarily disabled due to a Databricks-set rate limit of 0.

indicates that access to the selected foundation model endpoint is currently blocked by a Databricks-managed quota or entitlement setting rather than by the Azure Databricks workspace SKU itself.

A few points to verify:

Confirm that the workspace is attached to a paid Azure subscription and that billing is fully enabled.

Verify that AI/Model Serving features and Foundation Models are supported and enabled for your workspace region.

Review the Model Serving and Foundation Model entitlement settings in your Databricks account console, if available.

Check whether other foundation models can be invoked successfully from the same workspace, or whether the issue is specific to Claude-hosted endpoints.

Because the reported rate limit is explicitly being returned by Databricks, Microsoft Learn support and community moderators do not have visibility into the Databricks-managed trust tier, entitlement status, or model-specific quotas assigned to an individual workspace. We therefore cannot confirm whether the workspace is on a restricted trust tier or manually enable Claude model quotas.

Answer 1

Winston Martinez 0

Hi @Charan Venkat Sai Thambi
I already opened case 2609010030004708, but the responses so far have only repeated general troubleshooting steps and have not addressed the underlying backend issue.

My paid Premium Serverless workspace in East US returns the following error for every tested GPT and Claude model:

PERMISSION_DENIED: The endpoint is temporarily disabled due to a Databricks-set rate limit of 0.

I have already confirmed that billing is active, the workspace is not on a free trial, the region supports Model Serving, and the required features and entitlements are enabled.

Please increase the case priority and escalate it to the Azure Databricks/Databricks backend engineering team to:

Confirm the workspace’s trust tier and model-serving entitlement.

Verify why its Databricks-managed rate limit is set to zero.

Remove or correct the zero quota.

Confirm eligibility for GPT and Anthropic Claude foundation models.

This issue is blocking all Foundation Model API usage. Please provide a substantive technical update and escalation status rather than repeating the standard troubleshooting guidance.

Thank you.

Charan Venkat Sai Thambi 0 Reputation points Microsoft External Staff Moderator

2026-09-03T09:13:37.47+00:00

Hello @Winston Martinez

Thank you for your response.

Let me check with the backend team and get back to you with an update.
Charan Venkat Sai Thambi 0 Reputation points Microsoft External Staff Moderator

2026-09-04T08:22:47.16+00:00

Hello @Winston Martinez

I got update from Yao Siabi that they have involved the relevant Databricks Accounts team to support the steps needed to upgrade your Databricks account.

The Accounts team will now review the request and provide the guidance needed to move forward with the account upgrade.
Charan Venkat Sai Thambi 0 Reputation points Microsoft External Staff Moderator

2026-09-10T10:24:59.7366667+00:00

Hello @Winston Martinez

Greetings of the day!

I wanted to check if there is any update on this ?

Answer 2

Winston Martinez 0

Thank you for your response. I have already verified the following:

Workspace mypersoulases

is attached to a paid Azure subscription, not a free trial, and billing is enabled and i also used Databricks-premium plan (not free trails)

It is a Premium Serverless workspace in East US, where AI Model Serving and Foundation Models are supported.

Microsoft Entra ID authentication and the required AI features are enabled.

I reviewed the available Model Serving and Foundation Model entitlement settings in the Databricks account console.

The issue affects all tested GPT and Claude models, so it is not specific to one endpoint.

Every request returns:

PERMISSION_DENIED: The endpoint is temporarily disabled due to a Databricks-set rate limit of 0.

This indicates that the workspace has been assigned a zero quota by a Databricks-managed backend entitlement or trust-tier restriction. I have exhausted the troubleshooting options available from my side.

Please escalate this to the appropriate Azure Databricks/Databricks backend team to:

Confirm the workspace’s current trust tier and model-serving entitlement.
Confirm the Databricks-managed quota currently assigned to it.
Remove the zero-rate limit or move the workspace to a fully enabled payable tier.
Enable access to the supported GPT and Claude foundation models.

Charan Venkat Sai Thambi 0 Reputation points Microsoft External Staff Moderator

2026-09-02T12:57:07.91+00:00
Hello @Winston Martinez ,

Thank you for the detailed information and for outlining the extensive troubleshooting you have already completed. I understand the frustration this issue may be causing, especially after confirming that the workspace is associated with a paid subscription, is running on the Premium tier with Serverless enabled, and that all available troubleshooting steps have been exhausted.

Based on the information provided, a deeper investigation by the Azure Databricks support team is required to validate the backend entitlement and quota configuration associated with your workspace.

To ensure this is reviewed by the appropriate team, I kindly request that you create a support request by following the document below:

https://learn.microsoft.com/en-us/azure/databricks/admin/account-settings/serverless-quotas#increase-quota

Please request the Azure Databricks team to:

Confirm the workspace's current trust tier and model-serving entitlement.

Verify the Databricks-managed quota currently assigned to the workspace.

Investigate and, if applicable, remove the zero-rate limit or update the workspace to the appropriate fully enabled tier.

Confirm eligibility and enablement for the supported GPT and Anthropic Claude foundation models.

We appreciate your patience and cooperation while this issue is reviewed by the backend engineering team. Creating the support request will ensure that the necessary Databricks-managed configurations and entitlements can be examined and addressed by the appropriate specialists.

Please let us know if you need any assistance while submitting the support request.