Managing Traffic Jams with Azure OpenAI PTU Spillover

I tried using the PTU Spillover feature by adding x-ms-spillover-deployment in the header through "by request". However, it did not work correctly. I still received a 429 response and was not automatically redirected to the standard deployment.

import requests
import json

url = "$AZURE_OPENAI_ENDPOINT/openai/deployments/{ptu-deployment}/chat/completions?api-version=2025-02-01-preview"

payload = json.dumps({
  "messages": [
    {
      "content": "Does Azure OpenAI support customer managed keys?",
      "role": "user"
    }
  ],
  "model": "gpt-4o",
  "stream": False
})
headers = {
    "Content-Type": "application/json",
    "x-ms-spillover-deployment": "{spillover-standard-deployment}",
    "Authorization": "Bearer YOUR_AUTH_TOKEN"
}

response = requests.request("POST", url, headers=headers, data=payload)

print(response.text)

I believe I have met the following prerequisites:
1. A primary (provisioned) deployment. (Model: Gpt-4o, Provisioned Managed)
2. A standard deployment (Model: Gpt-4o, Global Standard)
3. Both deployments need to be part of the same Azure OpenAI Service resource. (Deploy on same region)

The only part I’m unsure about is whether a provisioned deployment can spill over to a global standard deployment?
According to the documentation, both deployments must have the same data processing type (e.g., a global provisioned lane can only spill over to a global standard lane).

By the way, is there a minimum API version requirement for PTU Spillover?