-
I'm building a WhatsApp agent with RubyLLM. After a few questions asked and answered the flow just stops because of the error:
I get a set of data after a few shots (sometimes it happens in the second attempt). I don't know why this is hapenning Get data
Data
I already try to add manually the |
Beta Was this translation helpful? Give feedback.
Replies: 5 comments 9 replies
-
Do you have a minimal reproducible example? |
Beta Was this translation helpful? Give feedback.
-
I also have seen these popping up from time to time and it blocks the whole chat from continuing. I'd like to understand why this is considered a blocking error. In my case, it may happen more often when there is a timeout on a long response, but I haven't nailed it down yet. |
Beta Was this translation helpful? Give feedback.
-
Hi I have the same problem. It doesn't happen locally (in dev) but I do get these errors on staging, where the DB is slower. I can verify the tool was called and at least in my case I can trace back the error message to a message that is in the chat: Message 102 requested tool call (responderai-rails):41:in ` ': An assistant message with 'tool_calls' must be followed by tool messages responding to each 'tool_call_id'. The following tool_call_ids did not have response messages: call_EbOVE6YZUSiRVM9v3rARxwXr (RubyLLM::BadRequestError)[#<ToolCall:0x00007fed0c15e888 responderai-rails(prod)> Chat.find(18).messages.find(102) responderai-rails(prod)> Chat.find(18).messages.find(103) |
Beta Was this translation helpful? Give feedback.
-
The one I'm getting is this: RubyLLM::BadRequestError: messages.130: Different error because I'm on anthropic. But probably caused by the same race condition. It started after I switched to the rails integration for my chat interface a few days ago and has affected 5 users so far. On my list to look into. |
Beta Was this translation helpful? Give feedback.
-
This is fixed in the latest trunk. |
Beta Was this translation helpful? Give feedback.
This is fixed in the latest trunk.