Using an condeact agent that triggers multiple api calls to the code interpreters and LLM, a single invoke call could take dozens of seconds to the agent to finalize its response.
In this scenario, there is a panicing issue from the router.
From the logs, we can see that the agent itself makes multiple successful calls to the code interpreters through the router; however, an issue likely arises when returning the result from the agent to the user.
Related logs from the agent:
