From 96d4466a8f487bf9837e72dc335ae34c9c67558f Mon Sep 17 00:00:00 2001
From: Krrish Dholakia <krrishdholakia@gmail.com>
Date: Mon, 12 Aug 2024 11:29:49 -0700
Subject: [PATCH] fix(azure.py): return response headers acompletion +
 acompletion w/ streaming

Fixes https://github.com/BerriAI/litellm/issues/2440#issuecomment-2283259187
---
 litellm/llms/azure.py | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/litellm/llms/azure.py b/litellm/llms/azure.py
index 9df67c26e1c..876d3b899a7 100644
--- a/litellm/llms/azure.py
+++ b/litellm/llms/azure.py
@@ -745,6 +745,8 @@ class AzureChatCompletion(BaseLLM):
             return convert_to_model_response_object(
                 response_object=stringified_response,
                 model_response_object=model_response,
+                hidden_params={"headers": headers},
+                _response_headers=headers,
             )
         except AzureOpenAIError as e:
             ## LOGGING
@@ -905,6 +907,7 @@ class AzureChatCompletion(BaseLLM):
                 model=model,
                 custom_llm_provider="azure",
                 logging_obj=logging_obj,
+                _response_headers=headers,
             )
             return streamwrapper  ## DO NOT make this into an async for ... loop, it will yield an async generator, which won't raise errors if the response fails
         except Exception as e: