Hi, I have common configurations for a program that I want to share across my stacks. I don't want to repeat the same configurations so what we're thinking of doing is creating a common config file that our pulumi program reads in first which can be overridden by values in the pulumi.[stack].yaml file. I would like to have these common configs show up configuration entries in the pulumi console but the pulumi.Config api doesn't have any methods to do set/create configs. I believe the automation api has a way to do this but i'm not sure if this would be the right approach. any suggestions?

If you wanted to use Pulumi config, you could have a project / stack that was just configuration, output all the config values and then consume them as stack references in other projects/stacks

what if I want to have common config within 1 service (common to all stacks but specific to 1 service). I would like to keep this common config inside the service, not in separate project

Yes, that would work but I was hoping that I could do this without creating a new project/stack for each project. I also took a closer look at the automation api and doesn't look like I can use this when using the pulumi cli.

The other way would be to make use of the fact you're writing code and have a package that had a load of constants in it that you could share

that's an option we've considered for configurations that go across all projects. what would be nice is to be able to see these configs in the pulumi console for auditing, debugging, etc.. i guess there isn't a way to currently do this?

I would always suggest creating an issue if we're not allowing you do what you want, and this isn't the first time I've spoken about cross-stack config

i believe there is already an issue about this: https://github.com/pulumi/pulumi/issues/2307 and one for configs in general: https://github.com/pulumi/pulumi/issues/4604 but i will create one specifically for being able to set configs from the Config class.

You can set config in the config class for the stack you're in

but i don't think you can use this api at the same time you runnig your program from the pulumi cli?

You can run pulumi programs through the automation api, but you're right about the CLI

Hi Piers, it looks like pulumi insists on presents of Pulumi.stack.yaml ( for example Pulumi.dev.yaml) be present when 'pulumi up' called. We want to generate this file at runtime. I tried to create this file as a first step the pulumi script, before calling new pulumi.Config();, but it's already too late, it picks values already there. Is there a way to enforce pulumi to reload values before continue with deployment?

You can have an empty stack config file, there doesn't have to be anything in it

What we wanted to do is dynamically create the Pulumi.dev.yaml. Our pulumi program reads in a config.common.yaml and a config.dev.yaml files and then merges these two files together and then writes it to Pulumi.dev.yaml. We then read the configs by calling new pulumi.Config(); We were hoping it would read the new stack config file we created but it looks the Config class uses the stack configs that were present when pulumi up was called and ignores the file that we created.

here is more details:

Have you looked at Automation API? I don't think what you're asking is possible outside of that.

Thanks. I will check

Using the automation API you lose all the features of the CLI as well as all the third party integrations. I found this issue that documents some of the shortcomings: https://github.com/pulumi/pulumi/issues/5538 and suggests a solution: https://github.com/pulumi/automation-api-examples/tree/main/go/inline_local_hybrid which we will look at.