Hi 👋! I’m new to Pulumi but I’ve generally found the onboarding experience and docs to be great! However, I’ve got a little stuck setting up continuous delivery. Some context: • We have an app that historically has been mostly serverless in nature. • We’re currently using AWS amplify to codegen and deploy cloud formation stacks. • For a few (off topic) reasons we’re looking move away from amplify and bring in some more container based services I’m currently trying out Pulumi and trying to set up CI/CD to match our current workflow. The main sticking point is “Preview Builds” (ie stacks that are created when you open a PR and destroyed when you close it) Two Questions 1. I was a little surprised that this pattern wasn’t covered in the blogs or the docs. Is there a reason for this? 2. It looks like I’d have to do some scripting to get things to work. Is this approach (medium article) still valid?

i quickly skimmed thru the github workflow yaml and overall approach - having spent a lot of time in the past few weeks on github actions and pulumi - the answer to your second question is yes

Great, thanks @many-yak-61188! It looks like a pretty sane approach. My best guess for why it isn’t officially endorsed or catered for with the CLI would just be that

pulumi

has a broader use-case than the serverless frameworks. Preview builds is an obvious approach for most projects that use

serverless

, Amplify, Netlify, Vercel etc, given it generally doesn’t cost you anything to spin up a fresh stack. I guess users of

pulumi

are more likely to have stacks that you couldn’t just create duplicates of at will and therefore using

pulumi preview

to check the changes is the next best thing (to creating a fresh isolated environment for each PR).