From b0ba3b2bc443a58dd9e6bba133a2c1a7273f91cf Mon Sep 17 00:00:00 2001
From: xuan <118178092+yuee3@users.noreply.github.com>
Date: Thu, 29 Oct 2026 16:03:04 +0800
Subject: [PATCH 1/3] fix(streaming): move sync max-duration check inside try
 so Timeout flows through the failure pipeline

---
 litellm/litellm_core_utils/streaming_handler.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/litellm/litellm_core_utils/streaming_handler.py b/litellm/litellm_core_utils/streaming_handler.py
index 5542249d2c7..86b6bb854b3 100644
--- a/litellm/litellm_core_utils/streaming_handler.py
+++ b/litellm/litellm_core_utils/streaming_handler.py
@@ -2068,8 +2068,8 @@ class CustomStreamWrapper:
         cache_hit = False
         if self.custom_llm_provider is not None and self.custom_llm_provider == "cached_response":
             cache_hit = True
-        self._check_max_streaming_duration()
         try:
+            self._check_max_streaming_duration()
             if self.completion_stream is None:
                 self.fetch_sync_stream()
 

From 5d2706fe2a13ba83889ff3e5a12c83c10f55b3ab Mon Sep 17 00:00:00 2001
From: xuan <118178092+yuee3@users.noreply.github.com>
Date: Thu, 29 Oct 2026 16:03:41 +0800
Subject: [PATCH 2/3] test(streaming): assert sync max-duration Timeout
 dispatches failure handler and closes stream

---
 .../test_max_streaming_duration.py                | 15 +++++++++++++++
 1 file changed, 15 insertions(+)

diff --git a/tests/unit/litellm_core_utils/test_max_streaming_duration.py b/tests/unit/litellm_core_utils/test_max_streaming_duration.py
index 234bc83e3c3..e645d410419 100644
--- a/tests/unit/litellm_core_utils/test_max_streaming_duration.py
+++ b/tests/unit/litellm_core_utils/test_max_streaming_duration.py
@@ -65,6 +65,21 @@ class TestCustomStreamWrapperMaxDuration:
             with pytest.raises(litellm.Timeout):
                 wrapper.__next__()
 
+    def test_sync_next_timeout_dispatches_failure_handler_and_closes(self):
+        """The sync Timeout must flow through the failure pipeline (parity with __anext__)."""
+        wrapper = _make_custom_stream_wrapper()
+        wrapper.logging_obj.failure_handler = MagicMock()
+        wrapper._stream_created_time = time.time() - 20
+        with patch("litellm.constants.LITELLM_MAX_STREAMING_DURATION_SECONDS", 10.0):
+            with pytest.raises(litellm.Timeout):
+                wrapper.__next__()
+        assert wrapper._closed is True
+
+        # failure_handler is dispatched on a thread; give it a moment
+        deadline = time.time() + 2.0
+        while time.time() < deadline and wrapper.logging_obj.failure_handler.call_count == 0:
+            time.sleep(0.01)
+        wrapper.logging_obj.failure_handler.assert_called_once()
     @pytest.mark.asyncio
     async def test_should_raise_on_async_anext_when_exceeded(self):
         """__anext__ checks the limit before iterating and schedules failure logging."""

From 7e82c048e4183df947ad1842e4eca997cded6caa Mon Sep 17 00:00:00 2001
From: xuan <118178092+yuee3@users.noreply.github.com>
Date: Thu, 29 Oct 2026 16:04:08 +0800
Subject: [PATCH 3/3] style: add missing blank line before next test

---
 tests/unit/litellm_core_utils/test_max_streaming_duration.py | 1 +
 1 file changed, 1 insertion(+)

diff --git a/tests/unit/litellm_core_utils/test_max_streaming_duration.py b/tests/unit/litellm_core_utils/test_max_streaming_duration.py
index e645d410419..ddaf5868bec 100644
--- a/tests/unit/litellm_core_utils/test_max_streaming_duration.py
+++ b/tests/unit/litellm_core_utils/test_max_streaming_duration.py
@@ -80,6 +80,7 @@ class TestCustomStreamWrapperMaxDuration:
         while time.time() < deadline and wrapper.logging_obj.failure_handler.call_count == 0:
             time.sleep(0.01)
         wrapper.logging_obj.failure_handler.assert_called_once()
+
     @pytest.mark.asyncio
     async def test_should_raise_on_async_anext_when_exceeded(self):
         """__anext__ checks the limit before iterating and schedules failure logging."""