AI Search Endpoint Quota in Azure

[TX-Aggie-00]

I have seen the common responses to most in this thread, so I will get ahead of it by saying that our workspace is not a trial, but a premium workspace in Azure.  I am a global admin in the workspace as well.  It seems like a lot of the AI settings got choked by the recent June update

I have a single Azure subscription and 2 Databricks workspaces: 1hybrid, 1 serverless.  In the hybrid workspace, I have created a single AI Search endpoint with a single index.  I am now creating a Knowledge Assistant for the AI Search endpoint and I am getting the error:

RESOURCE_EXHAUSTED: RESOURCE_EXHAUSTED: QUOTA_EXCEEDED: Maximum number of AI Search endpoints per workspace exceeded quota of 1. [Request ID: c5164f0c-37dc-4f7a-9fd3-b0fa4a8578f9, Trace ID: 1ddbe0465dfd468d7852711dd2fd4c9a]

In Azure, there is no way to request a quota increase for this, because it does not exist.  It keeps directing me to increase compute.  I have sent an email to databricks support and have received no response after 2 days.  It is a brand new account, so we have no AE yet.  I am pretty much stuck with no one to turn to to get this fixed.  This must have been rolled-out and implemented in the early-June update.

Has anyone experienced this, and if so, how did you get this resolved.  We are migrating a client and we are pretty much dead in the water right now.

Thanks,

Scott

[balajij8]

Hi,

The Knowledge Assistant creation flow tries to provision its own dedicated AI Search endpoint behind the scenes alongside reusing the index you configured during the flow (selected existing AI Search index in the flow). 

The endpoint quota is generally a higher number for AI Search endpoints per workspace in paid workspaces. The hard cap of exactly 1 is unusual as its mostly for free editions.

You can delete the existing vector search index and try creating Knowledge Assistant with the Files in a Volume as the input source (instead of vector search index) and allow Knowledge Assistant manage (via internal search index) it internally.

You can try the same setup of Knowledge Assistant & Vector Search Index in the other workspace if feasible.

Please ensure that the AI Search index uses any of the embedding models databricks-gte-large-en or databricks-bge-large-en or databricks-qwen3-embedding-0-6b.

[TX-Aggie-00]

Yes, that was my understanding as well, which is fine since the endpoint it stands up will not incur costs because the KA AI Search endpoint will be empty (the index already has an endpoint we are paying for). I will not delete the index, since the delta process has been running for it and do not want to repay for the embedding (it is a fairly large index).  I just need to understand why the quota exists on the account and how to get it raised.

Thanks,

Scott

[balajij8]

You can open a support ticket in the Azure end with Databricks workspace resource, Search Index & other issue details.

[TX-Aggie-00]

Tenant only has basic support, so the only option is Community, get help plan, or request quota increase (which AI Search Endpoint quota option does not even exist).  This seems like a glitch on the Databricks side though.  

Appreciate your responses!

Scott

[balajij8]

You can request additional quota increases by emailing supportops@databricks.com & help@databricks.com

• Subject line - "AI Search endpoint quota increase request"
• Include workspace ID, region, requested limit, screenshot and justification

You can reference these points in it

• Documentation states limit should be 500 per workspace. More details here
• Workspace shows limit of 1 (which matches Free Edition, not Premium)
• You're on a Premium tier workspace in Azure
• Error Request ID - c5164f0c-37dc-4f7a-9fd3-b0fa4a8578f9
• Trace id - 1ddbe0465dfd468d7852711dd2fd4c9a
• Workspace ID - 

Ask them to investigate if your workspace was misconfigured with Free Edition quotas after the June update

[TX-Aggie-00]

Awesome, thanks for this Balajij8!  Email sent, so we will see what happens.

Thanks
Scott

[balajij8]

You can share the account ID, workspace IDs, and a screenshot or the full error in it.

[iyashk-DB]

Hi, this doesn't appear to be a product limitation.

The documented default quota is 500 AI Search endpoints per workspace, and the documentation explicitly notes that this limit can be increased through your account team. Because of that, a Premium workspace reporting a quota of 1 is much more consistent with a workspace entitlement or provisioning issue than with an actual product limit.

I'd continue working with supportops/help@databricks.com, but since you don't currently have an assigned Account Executive, it's also worth escalating through your partner manager or partner support channel if that's available. Entitlement and quota issues often move more quickly through those channels than through a general support inbox.

[TX-Aggie-00]

Yes, it is very odd.  I have several clients running premium SKUs and have never run into this before.  This is the first workspace I have spun up since the June update, so not sure if that is causation or coincidence.  I did reach out to my AE for another account and he confirmed that it seems like a misconfiguration.  Also, because this tenant is new, there is no AE assigned so he can't directly escalate to the tech team.
Usually, I'd kill the workspace and fire up a new one, but the index has been running the initial sync for the past 3 days and do not want to lose that work or repay for the embedding (well, not coming out of my pocket, but still)

Thanks
Scott