Is there a way to auto-retry failed agent runs after an API error?

Community Discussion
1

Hey all,

I have an agent running on launch variables that executes daily. For the past two days, the OpenAI API was giving me errors and I couldn’t get any output. I had to manually run it today, and this time it worked fine.

Is there any option in MindStudio to make an AI agent automatically retry after a failed run (due to errors like API issues, etc.)? Or any recommended error handling or retry strategies for situations like this?

Thanks!

2

Hi @Kubi,

The Run Workflow block includes an option to automatically retry a sub-workflow if it encounters an error. However, you can’t auto-retry the entire agent unless it’s being triggered by another service.

If it fits your use case, adding a Run Workflow block should resolve this.