tfgen

tfgen

tfgen generates some Terraform configurations to make your life easier. All generated files are JSON and will have the file extension .tf.json. Please note that this is not a Terraform wrapper.

Installation

Download the latest release from GitHub Releases.

Setup

Create a .tfgen.yml file somewhere in your Terraform project's path.

---
# terraform s3 remote backend
# this will be used to generate a `backend.tf.json` for each stack
stack_regexp: "^([a-z]|[A-Z]|[0-9]|-)+$"

s3_backend:
  aws_region: us-west-2
  bucket: my-terraform-s3-backend-bucket
  dynamodb_table: my-terraform-dynamodb-table

# your stacks must have one of the following as a parent directory
environments:
  - production
  - staging
  - dev

tfgen will walk up the directory path in your cwd until it finds a valid .tfgen.yml file. This allows you to have separate config files for environments, applications, etc. It all depends on the directory structure of your Terraform project.

For example:

terraform
    ├── .tfgen.yml
    ├── production
    │   ├── myapp
    │   ├── redis
    │   ├── mysql
    │   ├── alb
    ├── staging
    │   ├── myapp
    │   ├── redis
    │   ├── mysql
    │   ├── alb

Usage

Adding a new stack:

cd production
mkdir new-stack
cd new-stack
tfgen init-stack

This will generate a backend.tf.json file

Adding a terraform_remote_state data source:

tfgen remote-state production/vpc

This will generate a remote_states.tf.json file.

Example:

{
  "data": {
    "terraform_remote_state": {
      "production_vpc": {
        "backend": "s3",
        "config": {
          "bucket": "my-terraform-s3-backend-bucket",
          "dynamodb_table": "my-terraform-dynamodb-table",
          "encrypt": true,
          "key": "stacks/production/vpc/terraform.tfstate",
          "region": "us-west-2"
        }
      }
}

One can make use of this in a Terraform template like this:

locals {
  vpc_id = "${data.terraform_remote_state.production_vpc.vpc_id}"
}