Hello. Couple of questions to get oriented. I am coming from Terraform. I have configured pulumi to use gcp bucket backend. But it also creates a local stack.yaml file. Why is this necessary. Can it be configured not to use local stack configs alsways use the bucket backend files. ? If so, how? The main reason is, I have tenant specific pulumi configs and having local configs are not realistic

Do you mean

Pulumi.<stackname>.yaml

? That's not the state file, it's the config file for that stack. It's where you put value that discriminate between stacks, which might be things like the region you're deploying to or the URL of some 3rd party service something connects to.

Adding to this: The Terraform equivalent would be the

variable.<env>.tfvars

files. You can retrieve the latest configuration of a stack through the CLI via

pulumi config refresh

, which can be helpful when you dynamically create stack configurations in your CI workflows.

Hi, what does it mean to get the latest configuration of a stack? Does it mean printing the values from the

Pulumi.<stackname>.yaml

to the console?

pulumi config refresh

updates your

Pulumi.<stackname>.yaml

file based on the state. For example, if your CI workflow created a new stack

pr-123

and you now want to work with it, you can run

pulumi stack select pr-123

and then

pulumi config refresh

on your machine to get the config file.

I was suspecting that but I wasn't sure. It's pulling the remote stack config to local config file. I guess that is only possible with Pulumi Cloud service. I don't this see this could work in my case where stack files are in AWS S3 and dynamic configuration set in CI/DC pipeline is lost anyway.

This works with a custom backend as well. The configuration is part of the state file, even when you set it dynamically.