ChatVertexAI InvalidArgument Error (Gemini 2.5 Flash Model)

[naveenSquareShift]

While making the LLM call from spring ai, with full prompt(system and user prompt) passed via .user(fullPrompt) facing issue like,

INVALID_ARGUMENT: Unable to submit request because it must include at least one parts field, which describes the prompt input.

This occurs because the payload for API end point is not constructed and passed properly.

[yaner-here]

Could you please provide a group of minimum reproduction steps?

[naveenSquareShift]

While making LLM call using the below code snippet if the response is empty and throwing MALFORMED_FUNCTION_CALL I try to retry the LLM call with same prompts, at that time I'm getting INVALID_ARGUMENT: Unable to submit request because it must include at least one parts field, which describes the prompt input.

Code Snippet:
chatClient.prompt()
.advisors(spec -> spec.param(ChatMemory.CONVERSATION_ID, conversationID))
.toolCallbacks(toolCallbackProvider.getToolCallbacks())
.user(Prompt)
.call()
.content();

@yaner-here Is there anything else need to be
shared? Thanks.