How I'm getting started. What might I do differently?

[enqack]

This post is meant to show how I'm getting started and to ask for feedback on what might I do differently, or even correctly.

I am in the process of starting up a company and it's infrastructure so the overall setup is quite small. I have basic needs for applications and an email system for them to use. Human email address are provided with Google Workspace but it's too costly and feature-full for robot usage.

Setup

In order to avoid file conflicts when pulling in upstream changes I have created a directory named wa under fast to provide a namespace to contain customizations.

Infrastructure design:

• Custom GCE image using NixOS
• Single server acting as a container host.
• Container workloads:
  • email (postfix and friends)
  • n8n
  • databases
• External managed DNS (godday)

My cloud-foundation-fabric repo design:

1. Forked GoogleCloudPlatform/cloud-foundation-fabric
2. Created copy of fork and made it private
3. Datasets for each stage were copied to fast/wa/stages/<stage>
4. terraform.tfvars files were created for each stage in fast/wa/evns/prod/<stage>
5. Configured factories_config in terraform.tfvars files to point to copied datasets
6. Edited yaml files to suite environment/deployment

Running a stage:

cd fast/stages/0-org-setup
terraform init
terraform plan -var-file="../../wa/envs/prod/0-org-setup/terraform.tfvars" -out="../../wa/envs/prod/0-org-setup/prod.tfplan"
terraform apply -var-file="../../wa/envs/prod/0-org-setup/terraform.tfvars" "../../wa/envs/prod/0-org-setup/prod.tfplan"

Status

I have applied stages 0-org-setup and 1-vpcsc. I ended up editing fast/wa/stages/0-org-setup/datasets/classic/defaults.yaml and fast/wa/stages/1-vpcsc/datasets/classic/access-levels/geo.yaml.

I'm currently researching yaml configuration files and what I need to change for each stage of the second section (stages prefixed with 2-).

Questions

My namespacing method seems solid and seems to be recommended. But I wonder if there are any alternatives to copying all dataset files besides using different paths pointing to unmodified upstream locations in the factories_config variable.

I don't have a firm grasp on the .config.yaml files yet. Are they any help with my previous question? Are they much different than a default.yaml file?

Is there existing code intended to be used with GCE image creation, specifically uploading an image? I'm currently using NixOS facilities (flakes) and external terraform code to accomplish that.

Ultimately I'd like to have a means to set the values needed for the datasets in a top-level yaml file or the terraform.tfvars files I'm using. Is there any preconceived methods to achieve that?

Thanks

I was originally going to write my own terraform modules to accomplish a basic VPC and compute engine deployment with most of the trimmings as I did a decade or so ago on AWS. Having the Cloud Foundation Fast method available has relieved me of a lot of code production. I greatly appreciate the work everyone involved has done.

Thank you for reading and for any feedback you have.