ChatVertexAI incompatible with pre-built ReAct agent when responses are blocked

[aabmass]

I'm using LangGraph's pre-built ReAct agent create_react_agent() along with ChatVertexAI. Occasionally I was seeing 400 responses from VertexAI:

Retrying langchain_google_vertexai.chat_models._completion_with_retry.<locals>._completion_with_retry_inner in 4.0 seconds as it raised InvalidArgument: 400 Unable to submit request because it must include at least one parts field, which describes the prompt input.

Stacktrace

 Traceback (most recent call last):
  File "/usr/local/google/home/aaronabbott/repo/opentelemetry-python-contrib/instrumentation-genai/opentelemetry-instrumentation-vertexai/examples/langgraph-chatbot-demo/.venv/lib/python3.13/site-packages/google/api_core/grpc_helpers.py", line 76, in error_remapped_callable
    return callable_(*args, **kwargs)
  File "/usr/local/google/home/aaronabbott/repo/opentelemetry-python-contrib/instrumentation-genai/opentelemetry-instrumentation-vertexai/examples/langgraph-chatbot-demo/.venv/lib/python3.13/site-packages/grpc/_interceptor.py", line 277, in __call__
    response, ignored_call = self._with_call(
                             ~~~~~~~~~~~~~~~^
        request,
        ^^^^^^^^
    ...<4 lines>...
        compression=compression,
        ^^^^^^^^^^^^^^^^^^^^^^^^
    )
    ^
  File "/usr/local/google/home/aaronabbott/repo/opentelemetry-python-contrib/instrumentation-genai/opentelemetry-instrumentation-vertexai/examples/langgraph-chatbot-demo/.venv/lib/python3.13/site-packages/grpc/_interceptor.py", line 332, in _with_call
    return call.result(), call
           ~~~~~~~~~~~^^
  File "/usr/local/google/home/aaronabbott/repo/opentelemetry-python-contrib/instrumentation-genai/opentelemetry-instrumentation-vertexai/examples/langgraph-chatbot-demo/.venv/lib/python3.13/site-packages/grpc/_channel.py", line 440, in result
    raise self
  File "/usr/local/google/home/aaronabbott/repo/opentelemetry-python-contrib/instrumentation-genai/opentelemetry-instrumentation-vertexai/examples/langgraph-chatbot-demo/.venv/lib/python3.13/site-packages/grpc/_interceptor.py", line 315, in continuation
    response, call = self._thunk(new_method).with_call(
                     ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^
        request,
        ^^^^^^^^
    ...<4 lines>...
        compression=new_compression,
        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
    )
    ^
  File "/usr/local/google/home/aaronabbott/repo/opentelemetry-python-contrib/instrumentation-genai/opentelemetry-instrumentation-vertexai/examples/langgraph-chatbot-demo/.venv/lib/python3.13/site-packages/grpc/_channel.py", line 1198, in with_call
    return _end_unary_response_blocking(state, call, True, None)
  File "/usr/local/google/home/aaronabbott/repo/opentelemetry-python-contrib/instrumentation-genai/opentelemetry-instrumentation-vertexai/examples/langgraph-chatbot-demo/.venv/lib/python3.13/site-packages/grpc/_channel.py", line 1006, in _end_unary_response_blocking
    raise _InactiveRpcError(state)  # pytype: disable=not-instantiable
    ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
grpc._channel._InactiveRpcError: <_InactiveRpcError of RPC that terminated with:
	status = StatusCode.INVALID_ARGUMENT
	details = "Unable to submit request because it must include at least one parts field, which describes the prompt input. Learn more: https://cloud.google.com/vertex-ai/generative-ai/docs/model-reference/gemini"
	debug_error_string = "UNKNOWN:Error received from peer ipv4:142.251.16.95:443 {grpc_message:"Unable to submit request because it must include at least one parts field, which describes the prompt input. Learn more: https://cloud.google.com/vertex-ai/generative-ai/docs/model-reference/gemini", grpc_status:3, created_time:"2025-04-10T20:45:21.468704714+00:00"}"
>

The above exception was the direct cause of the following exception:

Traceback (most recent call last):
  File "/usr/local/google/home/aaronabbott/repo/opentelemetry-python-contrib/instrumentation-genai/opentelemetry-instrumentation-vertexai/examples/langgraph-chatbot-demo/.venv/lib/python3.13/site-packages/opentelemetry/trace/__init__.py", line 587, in use_span
    yield span
  File "/usr/local/google/home/aaronabbott/repo/opentelemetry-python-contrib/instrumentation-genai/opentelemetry-instrumentation-vertexai/examples/langgraph-chatbot-demo/.venv/lib/python3.13/site-packages/opentelemetry/sdk/trace/__init__.py", line 1105, in start_as_current_span
    yield span
  File "/usr/local/google/home/aaronabbott/repo/opentelemetry-python-contrib/instrumentation-genai/opentelemetry-instrumentation-vertexai/examples/langgraph-chatbot-demo/src/langgraph_chatbot_demo/langchain_history.py", line 215, in <module>
    res = app.invoke({"messages": [message]}, config=config)
  File "/usr/local/google/home/aaronabbott/repo/opentelemetry-python-contrib/instrumentation-genai/opentelemetry-instrumentation-vertexai/examples/langgraph-chatbot-demo/.venv/lib/python3.13/site-packages/langgraph/pregel/__init__.py", line 2669, in invoke
    for chunk in self.stream(
                 ~~~~~~~~~~~^
        input,
        ^^^^^^
    ...<6 lines>...
        **kwargs,
        ^^^^^^^^^
    ):
    ^
  File "/usr/local/google/home/aaronabbott/repo/opentelemetry-python-contrib/instrumentation-genai/opentelemetry-instrumentation-vertexai/examples/langgraph-chatbot-demo/.venv/lib/python3.13/site-packages/langgraph/pregel/__init__.py", line 2323, in stream
    for _ in runner.tick(
             ~~~~~~~~~~~^
        loop.tasks.values(),
        ^^^^^^^^^^^^^^^^^^^^
    ...<2 lines>...
        get_waiter=get_waiter,
        ^^^^^^^^^^^^^^^^^^^^^^
    ):
    ^
  File "/usr/local/google/home/aaronabbott/repo/opentelemetry-python-contrib/instrumentation-genai/opentelemetry-instrumentation-vertexai/examples/langgraph-chatbot-demo/.venv/lib/python3.13/site-packages/langgraph/pregel/runner.py", line 146, in tick
    run_with_retry(
    ~~~~~~~~~~~~~~^
        t,
        ^^
    ...<11 lines>...
        },
        ^^
    )
    ^
  File "/usr/local/google/home/aaronabbott/repo/opentelemetry-python-contrib/instrumentation-genai/opentelemetry-instrumentation-vertexai/examples/langgraph-chatbot-demo/.venv/lib/python3.13/site-packages/langgraph/pregel/retry.py", line 40, in run_with_retry
    return task.proc.invoke(task.input, config)
           ~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/google/home/aaronabbott/repo/opentelemetry-python-contrib/instrumentation-genai/opentelemetry-instrumentation-vertexai/examples/langgraph-chatbot-demo/.venv/lib/python3.13/site-packages/langgraph/utils/runnable.py", line 600, in invoke
    input = step.invoke(input, config, **kwargs)
  File "/usr/local/google/home/aaronabbott/repo/opentelemetry-python-contrib/instrumentation-genai/opentelemetry-instrumentation-vertexai/examples/langgraph-chatbot-demo/.venv/lib/python3.13/site-packages/langgraph/utils/runnable.py", line 357, in invoke
    ret = context.run(self.func, *args, **kwargs)
  File "/usr/local/google/home/aaronabbott/repo/opentelemetry-python-contrib/instrumentation-genai/opentelemetry-instrumentation-vertexai/examples/langgraph-chatbot-demo/.venv/lib/python3.13/site-packages/langgraph/prebuilt/chat_agent_executor.py", line 686, in call_model
    response = cast(AIMessage, model_runnable.invoke(state, config))
                               ~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^
  File "/usr/local/google/home/aaronabbott/repo/opentelemetry-python-contrib/instrumentation-genai/opentelemetry-instrumentation-vertexai/examples/langgraph-chatbot-demo/.venv/lib/python3.13/site-packages/langchain_core/runnables/base.py", line 3025, in invoke
    input = context.run(step.invoke, input, config)
  File "/usr/local/google/home/aaronabbott/repo/opentelemetry-python-contrib/instrumentation-genai/opentelemetry-instrumentation-vertexai/examples/langgraph-chatbot-demo/.venv/lib/python3.13/site-packages/langchain_core/runnables/base.py", line 5358, in invoke
    return self.bound.invoke(
           ~~~~~~~~~~~~~~~~~^
        input,
        ^^^^^^
        self._merge_configs(config),
        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
        **{**self.kwargs, **kwargs},
        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
    )
    ^
  File "/usr/local/google/home/aaronabbott/repo/opentelemetry-python-contrib/instrumentation-genai/opentelemetry-instrumentation-vertexai/examples/langgraph-chatbot-demo/.venv/lib/python3.13/site-packages/langchain_core/language_models/chat_models.py", line 307, in invoke
    self.generate_prompt(
    ~~~~~~~~~~~~~~~~~~~~^
        [self._convert_input(input)],
        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
    ...<6 lines>...
        **kwargs,
        ^^^^^^^^^
    ).generations[0][0],
    ^
  File "/usr/local/google/home/aaronabbott/repo/opentelemetry-python-contrib/instrumentation-genai/opentelemetry-instrumentation-vertexai/examples/langgraph-chatbot-demo/.venv/lib/python3.13/site-packages/langchain_core/language_models/chat_models.py", line 843, in generate_prompt
    return self.generate(prompt_messages, stop=stop, callbacks=callbacks, **kwargs)
           ~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/google/home/aaronabbott/repo/opentelemetry-python-contrib/instrumentation-genai/opentelemetry-instrumentation-vertexai/examples/langgraph-chatbot-demo/.venv/lib/python3.13/site-packages/langchain_core/language_models/chat_models.py", line 683, in generate
    self._generate_with_cache(
    ~~~~~~~~~~~~~~~~~~~~~~~~~^
        m,
        ^^
    ...<2 lines>...
        **kwargs,
        ^^^^^^^^^
    )
    ^
  File "/usr/local/google/home/aaronabbott/repo/opentelemetry-python-contrib/instrumentation-genai/opentelemetry-instrumentation-vertexai/examples/langgraph-chatbot-demo/.venv/lib/python3.13/site-packages/langchain_core/language_models/chat_models.py", line 908, in _generate_with_cache
    result = self._generate(
        messages, stop=stop, run_manager=run_manager, **kwargs
    )
  File "/usr/local/google/home/aaronabbott/repo/opentelemetry-python-contrib/instrumentation-genai/opentelemetry-instrumentation-vertexai/examples/langgraph-chatbot-demo/.venv/lib/python3.13/site-packages/langchain_google_vertexai/chat_models.py", line 1310, in _generate
    return self._generate_gemini(
           ~~~~~~~~~~~~~~~~~~~~~^
        messages=messages,
        ^^^^^^^^^^^^^^^^^^
    ...<3 lines>...
        **kwargs,
        ^^^^^^^^^
    )
    ^
  File "/usr/local/google/home/aaronabbott/repo/opentelemetry-python-contrib/instrumentation-genai/opentelemetry-instrumentation-vertexai/examples/langgraph-chatbot-demo/.venv/lib/python3.13/site-packages/langchain_google_vertexai/chat_models.py", line 1546, in _generate_gemini
    response = _completion_with_retry(
        self.prediction_client.generate_content,
    ...<5 lines>...
        **kwargs,
    )
  File "/usr/local/google/home/aaronabbott/repo/opentelemetry-python-contrib/instrumentation-genai/opentelemetry-instrumentation-vertexai/examples/langgraph-chatbot-demo/.venv/lib/python3.13/site-packages/langchain_google_vertexai/chat_models.py", line 636, in _completion_with_retry
    return _completion_with_retry_inner(
        generation_method,
        **params,
    )
  File "/usr/local/google/home/aaronabbott/repo/opentelemetry-python-contrib/instrumentation-genai/opentelemetry-instrumentation-vertexai/examples/langgraph-chatbot-demo/.venv/lib/python3.13/site-packages/tenacity/__init__.py", line 336, in wrapped_f
    return copy(f, *args, **kw)
  File "/usr/local/google/home/aaronabbott/repo/opentelemetry-python-contrib/instrumentation-genai/opentelemetry-instrumentation-vertexai/examples/langgraph-chatbot-demo/.venv/lib/python3.13/site-packages/tenacity/__init__.py", line 475, in __call__
    do = self.iter(retry_state=retry_state)
  File "/usr/local/google/home/aaronabbott/repo/opentelemetry-python-contrib/instrumentation-genai/opentelemetry-instrumentation-vertexai/examples/langgraph-chatbot-demo/.venv/lib/python3.13/site-packages/tenacity/__init__.py", line 376, in iter
    result = action(retry_state)
  File "/usr/local/google/home/aaronabbott/repo/opentelemetry-python-contrib/instrumentation-genai/opentelemetry-instrumentation-vertexai/examples/langgraph-chatbot-demo/.venv/lib/python3.13/site-packages/tenacity/__init__.py", line 418, in exc_check
    raise retry_exc.reraise()
          ~~~~~~~~~~~~~~~~~^^
  File "/usr/local/google/home/aaronabbott/repo/opentelemetry-python-contrib/instrumentation-genai/opentelemetry-instrumentation-vertexai/examples/langgraph-chatbot-demo/.venv/lib/python3.13/site-packages/tenacity/__init__.py", line 185, in reraise
    raise self.last_attempt.result()
          ~~~~~~~~~~~~~~~~~~~~~~~~^^
  File "/usr/local/google/home/aaronabbott/.local/share/uv/python/cpython-3.13.1-linux-x86_64-gnu/lib/python3.13/concurrent/futures/_base.py", line 449, in result
    return self.__get_result()
           ~~~~~~~~~~~~~~~~~^^
  File "/usr/local/google/home/aaronabbott/.local/share/uv/python/cpython-3.13.1-linux-x86_64-gnu/lib/python3.13/concurrent/futures/_base.py", line 401, in __get_result
    raise self._exception
  File "/usr/local/google/home/aaronabbott/repo/opentelemetry-python-contrib/instrumentation-genai/opentelemetry-instrumentation-vertexai/examples/langgraph-chatbot-demo/.venv/lib/python3.13/site-packages/tenacity/__init__.py", line 478, in __call__
    result = fn(*args, **kwargs)
  File "/usr/local/google/home/aaronabbott/repo/opentelemetry-python-contrib/instrumentation-genai/opentelemetry-instrumentation-vertexai/examples/langgraph-chatbot-demo/.venv/lib/python3.13/site-packages/langchain_google_vertexai/chat_models.py", line 629, in _completion_with_retry_inner
    return generation_method(**kwargs)
  File "/usr/local/google/home/aaronabbott/repo/opentelemetry-python-contrib/instrumentation-genai/opentelemetry-instrumentation-vertexai/examples/langgraph-chatbot-demo/.venv/lib/python3.13/site-packages/opentelemetry/instrumentation/vertexai/patch.py", line 168, in generate_content
    response = wrapped(*args, **kwargs)
  File "/usr/local/google/home/aaronabbott/repo/opentelemetry-python-contrib/instrumentation-genai/opentelemetry-instrumentation-vertexai/examples/langgraph-chatbot-demo/.venv/lib/python3.13/site-packages/google/cloud/aiplatform_v1beta1/services/prediction_service/client.py", line 2396, in generate_content
    response = rpc(
        request,
    ...<2 lines>...
        metadata=metadata,
    )
  File "/usr/local/google/home/aaronabbott/repo/opentelemetry-python-contrib/instrumentation-genai/opentelemetry-instrumentation-vertexai/examples/langgraph-chatbot-demo/.venv/lib/python3.13/site-packages/google/api_core/gapic_v1/method.py", line 131, in __call__
    return wrapped_func(*args, **kwargs)
  File "/usr/local/google/home/aaronabbott/repo/opentelemetry-python-contrib/instrumentation-genai/opentelemetry-instrumentation-vertexai/examples/langgraph-chatbot-demo/.venv/lib/python3.13/site-packages/google/api_core/grpc_helpers.py", line 78, in error_remapped_callable
    raise exceptions.from_grpc_error(exc) from exc
google.api_core.exceptions.InvalidArgument: 400 Unable to submit request because it must include at least one parts field, which describes the prompt input. Learn more: https://cloud.google.com/vertex-ai/generative-ai/docs/model-reference/gemini [detail: "[ORIGINAL ERROR] generic::invalid_argument: Unable to submit request because it must include at least one parts field, which describes the prompt input. Learn more: https://cloud.google.com/vertex-ai/generative-ai/docs/model-reference/gemini [google.rpc.error_details_ext] { message: \"Unable to submit request because it must include at least one parts field, which describes the prompt input. Learn more: https://cloud.google.com/vertex-ai/generative-ai/docs/model-reference/gemini\" } 525003295 { 1 { 1 { 1 { 1: \"blade:cloud.ai.nl.llm.proto.service.languagemodelservice-prod-us-central1\" 2: \"LanguageModelService\" 3: \"GenerateMultiModal\" } 2 { 1: 3 2: \"generic\" 3: \"Unable to submit request because it must include at least one parts field, which describes the prompt input. Learn more: https://cloud.google.com/vertex-ai/generative-ai/docs/model-reference/gemini\" 5 { 1 { 2: 82012934 3 { 2: \"Unable to submit request because it must include at least one parts field, which describes the prompt input. Learn more: https://c...(length 1195)"
]
During task with name 'agent' and id 'e3877e3b-5f8b-c7c1-5e45-62f47f273458'

I debugged the problem and it happens If Gemini returns a blocked response (finishes with SAFETY reason) in an earlier interaction. LangGraph's checkpointer keeps a record of the blocked message which ChatVertexAI turns into an empty Content with no Parts in the RPC. Filtering out the blocked messages fixes the problem:

class PatchedChatVertexAI(ChatVertexAI):
    def _prepare_request_gemini(
        self, messages: list[BaseMessage], *args: Any, **kwargs: Any
    ) -> v1GenerateContentRequest | GenerateContentRequest:
        # Filter out any blocked messages with no content which can appear if you have a blocked
        # message from finish_reason SAFETY:
        messages = [
            message
            for message in messages
            if not message.response_metadata.get("is_blocked", False)
        ]
        return super()._prepare_request_gemini(messages, *args, **kwargs)

I'm happy to submit a fix.

[theo-davia]

Hey @aabmass,
I encountered the same problem, thanks for finding the issue that was causing it !