I’m attempting to set up an agent. I’m following the fundamental method outlined in this video:
https://www.youtube.com/watch?v=qIegseDJyIQ
I have a few tool workflows that I’ve created which interface with Salesforce to retrieve data. When I invoke the agent, it consistently times out. Executing the tool workflows independently is very fast, so that appears to be functioning correctly. There doesn't seem to be a method to inspect the agent's internal processing unless it successfully completes. It did function on a few occasions, but now it is not working at all.
I’ve pared down most of the agent instructions, and also removed all but one of the tools, but the issue persists. Is there any way to examine the agent during its execution to identify where it might be getting stuck?
I'm encountering the same issue. I've followed the same troubleshooting steps you did. I even tried changing the LLM API, but it didn't resolve the problem. I've contacted support and am waiting for a response.
I believe there's a timeout of approximately 3 minutes for the Agent to execute. You could experiment with a more straightforward process to observe if that makes any difference.
It’s as straightforward as I can make it. Running the only scenario it has access to takes approximately 2 seconds.
Just to circle back on this. I resolved this by simplifying my Slack integration. I had been creating a placeholder message to edit later, but once I removed that, it functioned correctly.