Share via

Cannot use endpoint created by Foundry or Azure OpenAI Service in my embedding skillset

Koenig, Jennifer 25 Reputation points
2026-02-13T11:16:46.6633333+00:00

I am creating an Azure Search with an embedding skillset. The skillset requires a resource URL with openai.azure.com. However when I create an Azure OpenAI Resource, I get the following endpoint:

https://germanywestcentral.api.cognitive.microsoft.com/

This endpoint is NOT ACCEPTED in my embedding skillset! The endpoint I get when creating a Foundry resource was also not accepted!

The Azure OpenAI Resource I created should give me this endpoint:

https://(my open ai service name).openai.azure.com

I am creating my resources with Terraform. On one deployment, the correct endpoint was created for Azure OpenAI but on the next deployment I got the regional endpoint. This is totally inconsistent! Also, if we're supposed to use the "new" Foundry, then its endpoints should **** well be accepted by the skillsets!

embedding-skillset-error

Azure OpenAI Service
Azure OpenAI Service
An Azure service that provides access to OpenAI’s GPT-3 models with enterprise capabilities.
0 comments No comments
{count} votes

2 answers

Sort by: Most helpful
  1. Koenig, Jennifer 25 Reputation points
    2026-02-13T13:40:36.5533333+00:00

    Here's the solution:

    when creating the Foundry or OpenAI Resource with Terraform, make sure to use the argument "custom_subdomain_name". It should be the same value as the "name" argument.


  2. Sridhar M 5,335 Reputation points Microsoft External Staff Moderator
    2026-02-13T12:28:29.2433333+00:00

    Hi Koenig, Jennifer

    You're very welcome! I’m glad the information was helpful.

    To summarize the resolution for others: when creating the Foundry or Azure OpenAI resource using Terraform, make sure to set the custom_subdomain_name argument and use the same value as the name argument. This ensures the correct *.openai.azure.com endpoint is created and resolves the issue.

    If you need any further assistance or run into issues during cleanup or redeployment, please don’t hesitate to reach out. I’ll be happy to help.

    Since I’ve converted comment into an answer, could you please take a moment to mark it as Accepted? This helps others in the community with the same question find the solution more easily.

    Thank you!

    0 comments No comments

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.