Handling LLM Errors in Hierarchical CrewAI Process with Callbacks

LLMs
1

I am running a CrewAI process in a hierarchical setup, but sometimes I encounter LLM errors during execution.

I would like to know:

  1. Is there a way to add a callback functionality that triggers automatically whenever an LLM error occurs?
  2. If yes, how can I implement it within the CrewAI framework?

Any guidance, best practices, or examples would be greatly appreciated.

Thanks in advance!

2

What do you want to achieve with the callback?

There isn’t anything like that as far as I know but if you can describe what you aim to achieve people might have better context to advise you on what to do.

3

Thanks for your response!

The main issue I am facing is that I occasionally get LLM errors like:

  • “Received None or empty response from LLM call.”
  • “An unknown error occurred. Please check the details below. Error details: Invalid response from LLM call - None or empty.”

If I run the same Crew with the same task multiple times, sometimes it works fine, but other times I get the above errors.

I want to handle these errors more gracefully—perhaps with a retry mechanism, a fallback response, or a way to detect and log these errors properly within the CrewAI framework.

Is there a recommended approach for handling such intermittent LLM failures? Would implementing a callback or exception handling mechanism help in this case?

Any guidance or best practices would be much appreciated!