Given the following minimal fogg.yml file:

version: 2
defaults:
  project: foo
  owner: foo@foobar.com
  terraform_version: 0.12.25
  backend:
    bucket: my-terraform-state-bucket
    dynamodb_table: terraform-statelocks
    profile: default
    region: us-west-2
accounts:
  logging-account: {}
  main-account: {}
envs:
  development:
    components:
      database: {}
      webserver: {}
modules:
  rds_instance: {}

fogg apply will generate the following directory structure.

All of the leaf directories that Fogg creates are terraform workspaces. Fogg installs fogg.tf files in each workspace to manage remote state data sources and state storage, as well as a Makefile for running fogg-provided make targets.

├── scripts
└── terraform
    ├── accounts
    │   ├── logging-account
    │   └── main-account
    ├── envs
    │   └── development
    │       ├── database
    │       └── webserver
    ├── global
    └── modules
        └── rds_instance

Fogg will also write an empty main.tf, variables.tf and outputs.tf file to each workspace to encourage best practices for workspace layout:

• Any workspaces that need to expose output values to other workspaces should define those outputs in outputs.tf
• main.tf can invoke terraform modules and define terraform-controlled resources
• Values that are local to a workspace go in a locals{} block in variables.tf