Deploy GCP Resources with CDK for Terraform(CDKTF)

Deploy GCP Resources with CDK for Terraform

What is Cloud Development Kit for Terraform (CDKTF)?

CDKTF allows users to produce Terraform configuration using programming languages such as C#, Python, TypeScript, Java, or Go. Users can use all Terraform providers and modules with CDKTF.

CDKTF uses the Cloud Development Kit from AWS, which provides a set of language-native frameworks for defining infrastructure, and adapters that let underlying provisioning tools use those definitions. 


 To use CDKTF, you need:

To follow this demo, you also need:

  • GCP Project

Install CDKTF

Use npm to install CDFTF like below:

npm install --global cdktf-cli@latest

see the installation command output below:

Verify the installation using “cdktf –help” command

cdktf terraform

Now we are going to create a GCP BigQuery dataset with CDKTF using Python.

Create and initialize the project for CDKTF

Let’s create directory for the project and navigate to the directory.

$ mkdir devops-counsel-cdktf-demo
$ cd devops-counsel-cdktf-demo/

Inside the directory, initialize CDKTF with Python template for Python language and the --local flag prevent CDKTF from using Terraform Cloud.

CDKTF prompts you for a project name and description. I used default values.

$ cdktf init --template=python --local

The above “cdktf init” command creates following files in project directory.

cloudshell:~/devops-counsel-cdktf-demo$ ls
cdktf.json help Pipfile

Now, install Google Cloud provider for CDKTF python package

pip install cdktf-cdktf-provider-google

Python code for creating a BigQuery Dataset.

Replace content with following code to create BigQuery dataset called cdkdemo_dataset

#!/usr/bin/env python
from constructs import Construct
from cdktf import App, TerraformStack, TerraformOutput
from import (

class MyStack(TerraformStack):
    def __init__(self, scope: Construct, ns: str):
        super().__init__(scope, ns)

        GoogleProvider(self, id="bigquery-tf", region="us-central1", project="devops-counsel-demo")

        cdkdemoBQDataset = BigqueryDataset(self, "cdkdemo", dataset_id="cdkdemo")
        TerraformOutput(self, "bigquery-dataset-id",

app = App()
MyStack(app, "devops-counsel-cdktf-demo")


Add google-cloud provider “google@~> 3.3” to terraformProviders list in cdktf.json file like below.

cloudshell:~/devops-counsel-cdktf-demo$ cat cdktf.json
  "language": "python",
  "app": "pipenv run python",
  "projectId": "8209d1e4-55b1-4c58-b165-abf13f3583b5",
  "sendCrashReports": "false",
  "terraformProviders": ["google@~> 3.3"],
  "terraformModules": [],
  "codeMakerOutput": "imports",
  "context": {
    "excludeStackIdFromLogicalIds": "true",
    "allowSepCharsInLogicalIds": "true"

Now run “cdktf get” to Generate CDK Constructs for GCP Terraform provider and modules, they will be stored under imports directory.

cloudshell:~/devops-counsel-cdktf-demo$ cdktf get
Generated python constructs in the output directory: imports
cloudshell:~/devops-counsel-cdktf-demo$ ls
cdktf.json  help  imports  Pipfile

Run “cdktf deploy” to deploy the dataset using the Python code in file

$ cdktf apply

you can see terraform state file after apply has been finished successfully.

cloudshell:~/devops-counsel-cdktf-demo$ ls
cdktf.json  cdktf.out  help  imports  Pipfile  terraform.devops-counsel-cdktf-demo.tfstate

Deleting BigQuery Dataset using destroy

Run “cdftf destroy” to delete the BigQuery dataset that we have created earlier.

$ cdktf destroy


In this demo, I have installed CDKTF cli and initialized CDKTF project and deployed a simple BigQuery dataset using Python language.

I have also covered few CDKTF cli commands.

You can find more information on CDKTF in terraform official documentation.

For more on Terraform:

Terraform: for_each and count meta-arguments

Leave a Reply