This message was deleted.

Yes, since Pulumi uses full programming languages, you can take advantage of the native abstraction capabilities of the language. Functions and classes are the most common solution, and you can bundle these up in your language’s package manager for other people to use.

So really the best workflow is to write a class library that would provision the resources I want/need, push this to a Nuget server, and then each of my stacks calls these? The key is, if there needs to be a change to the environment, this change is done in the consumed class library (e.g. provision another ec2 machine, etc), otherwise the change needs to be implemented per environment.

That’s right. You can parameterize the classes however makes sense, and then set the config per environment. There’s a lot of flexibility, but typical software architecture practices apply.

Yup. So in my Terraform analogy, the module would be the class library I write that's consumed by the stacks.