From 4d01b6428448225807e6605d04e37e29fe729b44 Mon Sep 17 00:00:00 2001
From: Software Developer <7852635+dsuhinin@users.noreply.github.com>
Date: Thu, 20 Nov 2025 21:00:33 +0100
Subject: [PATCH] [Bugfix] - Add Trace Headers to Beam Search Path (#29100)

Signed-off-by: dsuhinin <suhinin.dmitriy@gmail.com>
---
 vllm/entrypoints/openai/serving_chat.py       | 1 +
 vllm/entrypoints/openai/serving_completion.py | 1 +
 vllm/entrypoints/openai/serving_engine.py     | 2 ++
 3 files changed, 4 insertions(+)

diff --git a/vllm/entrypoints/openai/serving_chat.py b/vllm/entrypoints/openai/serving_chat.py
index 59e1c8d531793..6cc685acd6728 100644
--- a/vllm/entrypoints/openai/serving_chat.py
+++ b/vllm/entrypoints/openai/serving_chat.py
@@ -319,6 +319,7 @@ class OpenAIServingChat(OpenAIServing):
                         request_id=request_id,
                         params=sampling_params,
                         lora_request=lora_request,
+                        trace_headers=trace_headers,
                     )
                 else:
                     engine_request, tokenization_kwargs = await self._process_inputs(
diff --git a/vllm/entrypoints/openai/serving_completion.py b/vllm/entrypoints/openai/serving_completion.py
index a114b77ebc16b..1cfb45ef40366 100644
--- a/vllm/entrypoints/openai/serving_completion.py
+++ b/vllm/entrypoints/openai/serving_completion.py
@@ -216,6 +216,7 @@ class OpenAIServingCompletion(OpenAIServing):
                         request_id=request_id,
                         params=sampling_params,
                         lora_request=lora_request,
+                        trace_headers=trace_headers,
                     )
                 else:
                     engine_request, tokenization_kwargs = await self._process_inputs(
diff --git a/vllm/entrypoints/openai/serving_engine.py b/vllm/entrypoints/openai/serving_engine.py
index 127b8e6dcb87c..7dab5dbacd28c 100644
--- a/vllm/entrypoints/openai/serving_engine.py
+++ b/vllm/entrypoints/openai/serving_engine.py
@@ -343,6 +343,7 @@ class OpenAIServing:
         request_id: str,
         params: BeamSearchParams,
         lora_request: LoRARequest | None = None,
+        trace_headers: Mapping[str, str] | None = None,
     ) -> AsyncGenerator[RequestOutput, None]:
         beam_width = params.beam_width
         max_tokens = params.max_tokens
@@ -437,6 +438,7 @@ class OpenAIServing:
                             beam_search_params,
                             request_id_item,
                             lora_request=lora_req,
+                            trace_headers=trace_headers,
                         )
                     )
                 )