fix(azure/passthrough): populate standard_logging_object via logging hook

litellm/llms/azure/passthrough/transformation.py

@@ -1,7 +1,9 @@
 from typing import TYPE_CHECKING, List, Optional, Tuple
 
 import httpx
+from httpx import Response
 
+from litellm.litellm_core_utils.litellm_logging import Logging
 from litellm.llms.azure.common_utils import BaseAzureLLM
 from litellm.llms.base_llm.passthrough.transformation import BasePassthroughConfig
 from litellm.secret_managers.main import get_secret_str
@@ -11,6 +13,8 @@
 if TYPE_CHECKING:
     from httpx import URL
 
+    from litellm.types.utils import CostResponseTypes
+
 
 class AzurePassthroughConfig(BasePassthroughConfig):
     def is_streaming_request(self, endpoint: str, request_data: dict) -> bool:
@@ -83,3 +87,36 @@
         self, api_key: Optional[str] = None, api_base: Optional[str] = None
     ) -> List[str]:
         return super().get_models(api_key, api_base)
+
+    def logging_non_streaming_response(
+        self,
+        model: str,
+        custom_llm_provider: str,
+        httpx_response: Response,
+        request_data: dict,
+        logging_obj: Logging,
+        endpoint: str,
+    ) -> Optional["CostResponseTypes"]:
+        from litellm import encoding
+        from litellm.llms.openai.chat.gpt_transformation import OpenAIGPTConfig
+        from litellm.types.utils import ModelResponse
+
+        if "chat/completions" not in endpoint:
+            return None
+
+        openai_chat_config = OpenAIGPTConfig()
+
+        litellm_model_response: ModelResponse = openai_chat_config.transform_response(
+            model=model,
+            messages=[{"role": "user", "content": "no-message-pass-through-endpoint"}],
+            raw_response=httpx_response,
+            model_response=ModelResponse(),
+            logging_obj=logging_obj,
+            optional_params={},
+            litellm_params={},
+            api_key="",
+            request_data=request_data,
+            encoding=encoding,
+        )
+
+        return litellm_model_response

tests/test_litellm/llms/azure/passthrough/test_azure_passthrough_transformation.py

@@ -0,0 +1,97 @@
+import json
+import os
+import sys
+from unittest.mock import MagicMock
+
+import httpx
+
+sys.path.insert(0, os.path.abspath("../../../../.."))
+
+from litellm.llms.azure.passthrough.transformation import AzurePassthroughConfig
+from litellm.types.utils import ModelResponse
+
+
+def _azure_chat_completion_body():
+    return {
+        "id": "chatcmpl-abc123",
+        "object": "chat.completion",
+        "created": 1700000000,
+        "model": "gpt-4.1-mini-2025-04-14",
+        "choices": [
+            {
+                "index": 0,
+                "message": {
+                    "role": "assistant",
+                    "content": "Hello! How can I assist you today?",
+                },
+                "finish_reason": "stop",
+            }
+        ],
+        "usage": {
+            "prompt_tokens": 10,
+            "completion_tokens": 8,
+            "total_tokens": 18,
+        },
+    }
+
+
+def _make_httpx_response(body: dict) -> httpx.Response:
+    return httpx.Response(
+        status_code=200,
+        headers={"content-type": "application/json"},
+        content=json.dumps(body).encode("utf-8"),
+        request=httpx.Request(
+            "POST",
+            "https://example.openai.azure.com/openai/deployments/gpt-4.1-mini/chat/completions",
+        ),
+    )
+
+
+def test_azure_passthrough_logging_non_streaming_response_chat_completions():
+    """
+    Returns a populated ModelResponse (with usage + content) for a chat/completions
+    endpoint. This is what _success_handler_helper_fn needs to build
+    standard_logging_object — without it, Datadog/cost-tracking/router-success all
+    raise on every Azure passthrough request.
+    """
+    config = AzurePassthroughConfig()
+    logging_obj = MagicMock()
+
+    result = config.logging_non_streaming_response(
+        model="gpt-4.1-mini",
+        custom_llm_provider="azure",
+        httpx_response=_make_httpx_response(_azure_chat_completion_body()),
+        request_data={
+            "model": "gpt-4.1-mini",
+            "messages": [{"role": "user", "content": "hi"}],
+        },
+        logging_obj=logging_obj,
+        endpoint="openai/deployments/gpt-4.1-mini/chat/completions",
+    )
+
+    assert isinstance(result, ModelResponse)
+    assert result.choices[0].message.content == "Hello! How can I assist you today?"
+    assert result.usage.prompt_tokens == 10
+    assert result.usage.completion_tokens == 8
+    assert result.usage.total_tokens == 18
+
+
+def test_azure_passthrough_logging_non_streaming_response_unknown_endpoint_returns_none():
+    """
+    Endpoints other than chat/completions (responses, messages, images) fall
+    through to None — matches base-class behavior and Bedrock's "unknown
+    endpoint" handling. Not a regression; just scoping.
+    """
+    config = AzurePassthroughConfig()
+    logging_obj = MagicMock()
+
+    result = config.logging_non_streaming_response(
+        model="gpt-4.1-mini",
+        custom_llm_provider="azure",
+        httpx_response=_make_httpx_response(_azure_chat_completion_body()),
+        request_data={},
+        logging_obj=logging_obj,
+        endpoint="openai/responses",
+    )
+
+    assert result is None