Try asking in the
#CRFURDVQB channel and/or the
#C019YSXN04B as this channel gets a bit noisy and the folks who know Helm and Automation API might not see it. Generally, Automation API programs can be hooked into a standard debugger and run that way. We've got a few more docs here that explain how to get diagnostic data in general, and perhaps that might also help (e.g., setting the TF_LOG environment variable):
https://www.pulumi.com/docs/troubleshooting/#diagnosing-issues