No matter how you like to participate in developer communities, Pulumi wants to meet you there. If you want to meet other Pulumi users to share use-cases and best practices, contribute code or documentation, see us at an event, or just tell a story about something cool you did with Pulumi, you are part of our community.

Pulumi Community

Does using a `pulumi.RunError` work better?

Looks like it still kills the outputs:
```Diagnostics:
  pulumi:pulumi:Stack (auth-service-auth-service.development):
    error: Image version is not present!

Outputs:
  - dbConnectionString     : "[secret]"
  - debugDBConnectionString: "[secret]"```

That’s when I `throw new pulumi.RunError()` just doing a `new pulumi.RunError()` shows that things are succeeding and doesn’t exit the `up`

To clarify your requirements: you want to run all the code, including all the export statements that create your outputs, but you want to prevent any changes to your resources or state?

If that is the case, then you don't want to apply the up. You would want to preview then not run up, or run up and choose n.

Perhaps the thing to do is to invoke your Pulumi code from automation-api, and do those sorts of don't-even-start checks outside the call to up.

It’s part of our CI process so we run `pulumi up --refresh --skip-preview --non-interactive --yes`. The `IMAGE_VERSION` comes from a k8s Deployment so we set that in the ENV. If it isn’t present the `up` should fail but didn’t want to outputs to be deleted. I guess I should rethink the architecture of this. Probably better to check in the script that runs the `up` for the variable and put the control flow there instead of in the Pulumi code. Haven’t played around with the automation-api yet but that’s definitely on the list of things to do during our tooling upgrades.