CloudFormation to Terraform

This node CLI tool is used for generating both Terraform configuration files as well as Terraform state so that you can use Terraform to manage CloudFormation templates. To further clarify, it does not generate terraform configuration for the individual resources CloudFormation provisions, it generates an aws_cloudformation_stack resource so that you can manage your existing CloudFormation stacks with Terraform instead of or in conjunction with the AWS Console and CLI.

Getting Started

npm i -g @humanmade/cf-to-tf

Recommended Dependencies

As this was designed to generate Terraform resources, it'd be a good idea to install terraform. You can install the binary by itself or use a tool like brew to manage it for you.

It's also recommended to install json2hcl as this will assist in processing output from cf-to-tf later.

Demo

Let's use the following CloudFormation Stack response as an example:

{
    "Stacks": [
        {
            "StackId": "arn:aws:cloudformation:eu-central-1:123456789012:stack/foobarbaz/255491f0-71b8-11e7-a154-500c52a6cefe",
            "Description": "FooBar Stack",
            "Parameters": [
                {
                    "ParameterValue": "bar",
                    "ParameterKey": "foo"
                },
                {
                    "ParameterValue": "baz",
                    "ParameterKey": "bar"
                },
                {
                    "ParameterValue": "qux",
                    "ParameterKey": "baz"

                }
            ],
            "Tags": [
                {
                    "Value": "bar",
                    "Key": "foo"
                },
                {
                    "Value": "qux",
                    "Key": "baz"
                }
            ],
            "Outputs": [
                {
                    "Description": "Foobarbaz",
                    "OutputKey": "FooBarBaz",
                    "OutputValue": "output value"
                }
            ],
            "CreationTime": "2017-07-26T04:08:57.266Z",
            "Capabilities": [
                "CAPABILITY_IAM"
            ],
            "StackName": "foobarbaz",
            "NotificationARNs": [],
            "StackStatus": "CREATE_COMPLETE",
            "DisableRollback": true
        }
    ]
}

Running cf-to-tf --stack foobarbaz config | json2hcl | cf-to-tf clean-hcl | terraform fmt - will generate the following config:

resource "aws_cloudformation_stack" "network" {
  capabilities     = ["CAPABILITY_IAM"]
  disable_rollback = true
  name             = "foobarbaz"

  parameters = {
    foo = "bar"
    bar = "baz"
    baz = "qux"
  }

  tags = {
    foo = "bar"
    baz = "qux"
  }
}

Usage

Usage: cf-to-tf [options] [command]


Options:

  -s, --stack <stack>                 The CloudFormation stack to import
  -r, --resource-name <resourceName>  The name to assign the terraform resource
  -h, --help                          output usage information


Commands:

  config      Generates Terraform configuration in JSON
  state       Generates Terraform state file in JSON
  template    Prints the CloudFormation Stack Template
  clean-hcl   Cleans generated HCL according to my preferences

Terraform JSON Files

This tool is designed to be used in conjunction with other tools. It will only output the data to STDOUT and is designed to be piped to another program to write the file to a location. For example, to generate a configuration file for a stack named lambda-resources, we could do the following:

cf-to-tf -s lambda-resources config | tee main.tf.json

This command will fetch a CloudFormation stack named lambda-resources and generate the required Terraform configuration for it. We then pipe the output to tee which will write to a file named main.tf.json. Because HCL is JSON compatible, Terraform can read the main.tf.json natively.

Terraform State

To generate the associated Terraform state for this CloudFormation stack, you would run the following:

cf-to-tf -s lambda-resources state | tee terraform.tfstate

This will create a state file from scratch. It assumes you don't already have an existing state file in place. I'm considering updating the tool to write just the resource portion of the state so it can be added to an existing state file, but that wasn't an immediate priority.

Pretty Printing

Both of these commands will generate compressed JSON output, meaning whitespace has been stripped. To pretty print the output for enhanced readability, you could pipe the output to jq, and then to tee:

cf-to-tf -s lambda-resources config | jq '.' | tee main.tf.json
cf-to-tf -s lambda-resources state | jq '.' | tee terraform.tfstate

Generating HCL

It's also possible to use a tool called json2hcl to generate HCL:

cf-to-tf -s lambda-resources config | json2hcl | tee main.tf

Unfortunately, while json2hcl outputs valid HCL, it's not in the format I like. To solve that problem, the clean-hcl command is also available. To output HCL in the format you'll normally see, you can execute this chain:

cf-to-tf -s lambda-resources config | json2hcl | cf-to-tf clean-hcl | terraform fmt - | tee main.tf

We're doing the same thing we were doing before, but now we're also piping the result to cf-to-tf clean-hcl which formats the file a certain way, then piping it to terraform fmt - which formats the file further (primarily, this tool aligns = and adds newlines where necessary).

Reading from STDIN

It's also possible to have cf-to-tf read stack data from STDIN. For example, if you have the JSON response from the aws-cli call stored in a variable for re-use, you can do the following:

JSON="$(aws cloudformation describe-stacks --stack-name lambda-resources)"
echo "$JSON" | cf-to-tf -s - config

AWS Authentication

The command uses the AWS SDK under the hood to retrieve the CloudFormation stack details, so set your authentication credentials as you would normally (~/.aws/credentials, AWS_PROFILE, AWS_REGION, etc).

Batch Import scripts

For an example of how to use this script in batch operations importing multiple stacks in multiple regions, refer to this gist.