Structuring Infrastructure Code

InfraDots runs Terraform or OpenTofu; your repo is standard .tf (and optionally .tfvars, modules). Use required_version in a terraform {} block so everyone—and InfraDots—uses the same engine and version. See Terraform & OpenTofu basics for core concepts.

Small Infrastructure Footprints

A general rule in structuring your code we tend to recommend for small infrastructure footprints:

Single repo
Division into separate files by types

Example project layout:

my-infra/
├── main.tf          # Provider config and backend
├── iam.tf           # IAM roles and policies
├── s3.tf            # S3 buckets and policies
├── instances.tf     # EC2 instances
├── networking.tf    # VPC, subnets, security groups
├── variables.tf     # Input variable declarations
├── outputs.tf       # Output values
└── terraform.tfvars # Variable values (do not commit secrets)

A typical main.tf for this setup:

terraform {
  required_version = ">= 1.5.0"

  # InfraDots injects the remote backend for runs in the platform.
  # For local CLI use with InfraDots state, configure backend "remote" with your hostname/org/workspace.
  backend "remote" {
    organization = "my-org"
  }
}

provider "aws" {
  region = var.region
}

Large Infrastructure Footprints

Separate your code into workspaces for components that will generate a lot of resources. Further separation can be done per workspace and per environment:

infra/
├── bootstrap/           # Main workspace for bootstrapping
│   ├── main.tf
│   └── outputs.tf
├── iam/                 # IAM workspace - one per env
│   ├── main.tf
│   ├── roles.tf
│   └── policies.tf
├── storage/             # Buckets workspace - one per env
│   ├── main.tf
│   └── buckets.tf
└── compute/             # Compute workspace - one per env
    ├── main.tf
    ├── instances.tf
    └── autoscaling.tf

Buckets and bucket related policies — one workspace per env
IAM users and permissions — one workspace per env
Main workspace for bootstrapping the rest of the workspaces

💡 Tip

[!tip] Use the InfraDots Terraform provider to manage workspaces programmatically and keep your workspace configuration as code.