Quota Reduction Delays - GPT Datazone

Robert Janssen - Zenarate 0 Reputation points
2025-10-01T18:44:50.7933333+00:00

We have a Datazone Quota set to 9 million for an old model, and 1 million for a new model.

On going to the UI for the Resource, I reduce the quota to 110k. I click CONFIRM.

Expected value: I now have 8,890k to reallocate to the newer model.

Actual value: The old model STILL says 9 million, and the newer model can't allocate any of the quota.

Azure OpenAI in Foundry Models

1 answer

Sort by: Most helpful
  1. Anshika Varshney 15,710 Reputation points Microsoft External Staff Moderator
    2025-10-01T20:22:56.24+00:00

    Hello Robert Janssen - Zenarate,

    Thank you for reaching out on the Microsoft Q&A.

    When you reduce quota for an existing model, the changes don’t always reflect in the UI immediately. This usually happens because the backend takes some time to process and sync the updated values across resources.

    One thing you can try is refreshing and rechecking after a short while. In most cases, the updated quota becomes visible once the system completes processing. You may also want to verify the quota at the API level, since sometimes the API reflects the updated numbers earlier than the portal UI.

    Another possibility is that the newer model can’t take the reallocated quota yet because the old model’s quota reduction hasn’t fully propagated. This is more of a temporary synchronization delay than an actual issue with your setup.

    In short, this behavior is expected, and the quota should update after some time. Once it syncs, you should be able to allocate the freed-up quota to the newer model.

    I Hope this helps. Do let me know if you have any further queries.

    Thankyou!

    Was this answer helpful?

    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.