fix(proxy): extract spend tags from Starlette Headers on pass-through

Pass-through endpoints store request.headers as a Starlette Headers
Mapping, but _get_user_agent_tags / _get_extra_header_tags only accepted
dict, so request_tags stayed empty. Accept Mapping in both extractors.

Fixes #44032

litellm/litellm_core_utils/litellm_logging.py

@@ -6270,7 +6270,7 @@ class StandardLoggingPayloadSetup:
             return None
         user_agent_tags: list[str] | None = None
         headers: Final = proxy_server_request.get("headers", {})
-        if headers is not None and isinstance(headers, dict):
+        if headers is not None and isinstance(headers, Mapping):
             if "user-agent" in headers:
                 user_agent: Final = headers["user-agent"]
                 if user_agent is not None:
@@ -6295,7 +6295,7 @@ class StandardLoggingPayloadSetup:
             return None
 
         headers: Final = proxy_server_request.get("headers", {})
-        if not isinstance(headers, dict):
+        if not isinstance(headers, Mapping):
             return None
 
         header_tags: Final = []

tests/unit/litellm_core_utils/test_litellm_logging.py

@@ -2959,6 +2959,52 @@ def test_get_extra_header_tags():
                 delattr(litellm, "extra_spend_tag_headers")
 
 
+
+def test_header_derived_spend_tags_accept_starlette_headers_mapping():
+    """Pass-through stores Starlette Headers (a Mapping, not dict); tags must still extract.
+
+    Regression for https://github.com/BerriAI/litellm/issues/44032
+    """
+    import litellm
+    from starlette.datastructures import Headers
+
+    from litellm.litellm_core_utils.litellm_logging import StandardLoggingPayloadSetup
+
+    original_extra_headers = getattr(litellm, "extra_spend_tag_headers", None)
+    original_disable_ua = getattr(litellm, "disable_add_user_agent_to_request_tags", False)
+    try:
+        litellm.disable_add_user_agent_to_request_tags = False
+        litellm.extra_spend_tag_headers = ["x-custom-header"]
+        proxy_server_request = {
+            "headers": Headers(
+                {
+                    "user-agent": "curl/8.0.0",
+                    "x-custom-header": "tenant-a",
+                }
+            )
+        }
+
+        ua_tags = StandardLoggingPayloadSetup._get_user_agent_tags(proxy_server_request)
+        assert ua_tags is not None
+        assert "User-Agent: curl" in ua_tags
+        assert "User-Agent: curl/8.0.0" in ua_tags
+
+        extra_tags = StandardLoggingPayloadSetup._get_extra_header_tags(proxy_server_request)
+        assert extra_tags is not None
+        assert "x-custom-header: tenant-a" in extra_tags
+
+        request_tags = StandardLoggingPayloadSetup._get_request_tags(
+            litellm_params={},
+            proxy_server_request=proxy_server_request,
+        )
+        assert "User-Agent: curl" in request_tags
+        assert "x-custom-header: tenant-a" in request_tags
+    finally:
+        litellm.extra_spend_tag_headers = original_extra_headers
+        litellm.disable_add_user_agent_to_request_tags = original_disable_ua
+
+
+
 def test_response_cost_calculator_with_response_cost_in_hidden_params(logging_obj):
     from litellm import Router