What is the right way to handle "runtime" errors in a pulumi program/stack config/automation session? Right now when there's an exception thrown (through my own code or pulumi failing, config for example) stdout is spammed with unhandled errors

Resource monitor is terminating

and becomes a little tricker to spot the actual isue.

Is this what you're referring to? https://github.com/pulumi/pulumi/issues/8236

Exactly

I'd recommend upvoting the issue and adding a comment. It'll help with prioritization.