[Frontend] Improve StreamingResponse Exception Handling (vllm-project#11752)

Author: robertgshaw2-redhat

vllm/entrypoints/openai/serving_chat.py

@@ -301,7 +301,7 @@ async def chat_completion_stream_generator(
                 ] * num_choices
             else:
                 tool_parsers = [None] * num_choices
-        except RuntimeError as e:
+        except Exception as e:
             logger.exception("Error in tool parser creation.")
             data = self.create_streaming_error_response(str(e))
             yield f"data: {data}\n\n"
@@ -591,7 +591,7 @@ async def chat_completion_stream_generator(
                 completion_tokens=num_completion_tokens,
                 total_tokens=num_prompt_tokens + num_completion_tokens)
 
-        except ValueError as e:
+        except Exception as e:
             # TODO: Use a vllm-specific Validation Error
             logger.exception("Error in chat completion stream generator.")
             data = self.create_streaming_error_response(str(e))

vllm/entrypoints/openai/serving_completion.py

@@ -371,7 +371,7 @@ async def completion_stream_generator(
             # report to FastAPI middleware aggregate usage across all choices
             request_metadata.final_usage_info = final_usage_info
 
-        except ValueError as e:
+        except Exception as e:
             # TODO: Use a vllm-specific Validation Error
             data = self.create_streaming_error_response(str(e))
             yield f"data: {data}\n\n"