Get Started With AWS CDK

Create an AWS CDK workspace and deploy infrastructure with Harness IaCM.

Use this guide to create an AWS Cloud Development Kit (CDK) workspace. Deploy infrastructure through Harness Infrastructure as Code Management (IaCM).

What is IaCM?

IaC automates infrastructure management via code. IaCM enhances this by ensuring consistent, accountable, and repeatable deployments. Harness boosts IaCM with real-time cost estimation, automated policy enforcement, and drift detection, ensuring efficient provisioning and compliance with standards.

TERRAFORM TO OPENTOFU MIGRATION

Harness IaCM supports integration with all OpenTofu versions. For Terraform, we support all MPL versions up to 1.5.x, any BSL versions (from 1.6.0) are not supported.

Follow this OpenTofu migration guide to transition from Terraform to OpenTofu and leverage the benefits of this open-source alternative.

Go to What's Supported in IaCM for more information on supported Cloud Providers and Code/Git providers.

AWS Cloud Development Kit (CDK) lets you define cloud infrastructure using general-purpose programming languages. Harness IaCM runs your CDK workspaces and pipelines with Git-backed configuration, connectors, and a default deploy pipeline purpose-built for the CDK lifecycle.

This guide walks you from connectors and workspace creation through a standard CDK deploy pipeline (bootstrap, synth, diff, deploy) and an optional approval between diff and deploy.

What you will learn from this topic

Connectors and workspace: Create cloud and Git connectors, then create a CDK workspace configured with your language and package manager.
Deploy pipeline: Run a pipeline that executes the four CDK steps: bootstrap, synth, diff, and deploy.
Stack targeting: Optionally deploy a specific stack from your CDK app rather than the full application.
Plugin variables: Always set PLUGIN_AWS_REGION for the target AWS Region. Set PLUGIN_AWSCDK_STACKS only when you want specific stacks processed; otherwise all stacks in the app are included.
Approvals: Optionally gate deploy behind an approval step after diff.

Before you begin

Before you use this guide, ensure you have the following:

Harness account with IaCM enabled: You need Infrastructure as Code Management under Infrastructure in Harness when it is entitled on your account. Go to Getting started with Harness Platform to access or create a Harness account.
CONTACT HARNESS SUPPORT
If IaCM does not appear, contact your account administrator or Harness Support.
Pipeline permissions: You need View, Create/Edit, and Execute for Pipelines. Go to RBAC in Harness and Manage roles to configure roles.
AWS account: Access to an AWS account with permissions to deploy CloudFormation stacks. CDK bootstrapping requires permissions to create S3 buckets and ECR repositories in your target account and region.
Git repository: Access to a Git provider with your CDK project.
Harness organization and project: An organization and project set up on the Harness Platform.

Sample CDK app

The following example CDK app creates an S3 bucket and an SQS queue. Save it as app.py in the root of your repository, with a standard cdk.json pointing to it.

import aws_cdk as cdk
from aws_cdk import (
    aws_s3 as s3,
    aws_sqs as sqs,
    Stack,
)
from constructs import Construct

class MyFirstStack(Stack):
    def __init__(self, scope: Construct, construct_id: str, **kwargs) -> None:
        super().__init__(scope, construct_id, **kwargs)

        s3.Bucket(self, "MyBucket",
            versioned=True,
            removal_policy=cdk.RemovalPolicy.DESTROY,
        )

        sqs.Queue(self, "MyQueue",
            visibility_timeout=cdk.Duration.seconds(300),
        )

app = cdk.App()
MyFirstStack(app, "MyFirstStack")
app.synth()

Go to the AWS CDK documentation to review supported languages and CDK app structure.

Set up your workspace

A workspace is a named environment for storing your CDK configurations and resources. Connect your AWS account and code repository through Connectors to manage infrastructure changes and deployments with Harness IaCM pipelines.

CONNECTORS BEFORE WORKSPACE

Harness recommends configuring your connectors before you create a workspace. You can also add new connectors during the Create Workspace flow.

Create a connector

Use Harness AI to create and configure your AWS and code repository connectors before you create a workspace:

When you add a connector, start with:

Sign in to app.harness.io.
In the module pane, select Infrastructure.
Select Project Setup, and then select Connectors.
Select New Connector (AI).
Select an option, for example "Create an AWS connector", or type your request to create a connector for your AWS account or code repository.

Harness creates a YAML file for your connector. Once you select Create, Harness adds it to your project.

EDIT CONNECTOR

Edit your connector by updating the AI-generated YAML file, or by selecting Edit Details in the connectors panel.

Go to Connect your Cloud Provider to set up your cloud account connection. Go to Connect your Code Repository to set up your Git connection.

OIDC CONNECTORS

For easier access and token management, use the OIDC (OpenID Connect) option in the Credentials panel. This allows your connector to assume roles with permissions set in your AWS account, updated only by authorized users. Go to Use OIDC to review credential options.

Create your workspace

Once you have configured your connectors, create a workspace and select those connectors in the New Workspace wizard.

In the module pane, select Infrastructure.
Select an existing project or create a new project.
Select Workspaces, and then select New Workspace.
Select Create new Workspace, then select Start from scratch and complete the following fields:

About workspace

Name: Type a unique name to identify the workspace.
Description (optional): Type an optional description.
Tags (optional): Add tags to identify the workspace.

Configure repository details

Select your Git provider: Harness Code Repository or Third-party Git provider.
Git Connector: Select the Git connector you created in the previous step.
Git Fetch Type: Select Latest from branch, Git tag, or Commit SHA.
Git Branch: Specify the branch to use.
Folder Path: Specify the path from the repository root to the directory containing your CDK app (for example, src/my-cdk-app).

Provisioner

Connector: Select the AWS connector you created in the previous step.
Workspace Type: Select AWS CDK.
AWS CDK version: Select the CDK version to use.
Programming Language: Select the programming language for your CDK app (Python, TypeScript, JavaScript, Java, or Go).
Language Version: Select the language runtime version to use.
Package Manager: Select your package manager (for example, Pip for Python).
Package Manager Version: Select the package manager version.

DYNAMIC ENVIRONMENT INSTALLATION

Harness installs your selected language runtime and package manager dynamically at pipeline execution time rather than using pre-baked Docker images. This means each run uses exactly the versions you specify in the workspace configuration.

Select Create.

Run the deploy pipeline

CDK workspaces come with a default deploy pipeline that runs the four CDK lifecycle steps in sequence. Go to AWS CDK pipeline steps to review what each step does and how to configure it.

CDK plugin environment variables

Before you run the pipeline, set environment variables the CDK plugin expects:

PLUGIN_AWS_REGION (required): AWS Region where resources should be deployed (for example us-east-1). You always need this set explicitly, for example as a MANUAL row under Environment Variables on the workspace Connectors and Variables tab, or as stage, pipeline, or step variables.
PLUGIN_AWSCDK_STACKS (optional): Comma-separated logical stack IDs only when you want this pipeline run to target specific stacks (for example s3bucket or SQSStack,S3Stack). If you omit it, every stack in the app is processed. Harness is working on a more seamless experience for stack selection.

Add them as stage or pipeline variables, on the workspace Environment Variables table, or under Environment Variables on the CDK steps. Go to AWS Connector Authentication to review how other IaCM steps use the PLUGIN_* pattern. Go to Set up the AWS CDK provisioner to understand the CDK lifecycle and stacks.

Select the Infrastructure module.
Select Pipelines, then select Create a Pipeline.
Select an option from Harness AI chat or type a request to generate a CDK deploy pipeline, for example:
- "Create a CDK deploy pipeline with bootstrap, synth, diff, and deploy steps."
Review the generated YAML and Harness AI chat summary, then select Accept.
Select Run to execute the pipeline.

The pipeline runs four steps:

Bootstrap: Provisions the CDK toolkit resources (an S3 bucket and ECR repository) in your target AWS account and region. This step is required once per account and region before any CDK app can be deployed there.
Synth: Compiles your CDK app code into a CloudFormation template. This step validates your code and produces the CloudFormation assets used by subsequent steps.
Diff: Compares the synthesized CloudFormation template against the currently deployed stack and shows what will change.
Deploy: Executes the CloudFormation change set and provisions your infrastructure.

Target a specific stack (optional)

If your CDK app defines multiple stacks, you can target a single stack for deployment rather than running against the full app. Set the Stack Path field on your pipeline steps to the identifier of the stack you want to deploy.

WHEN TO USE STACK TARGETING

Use stack targeting when you have a monorepo CDK app with independent stacks per environment or service and you want to deploy only one at a time. Go to Set up the AWS CDK provisioner to understand how stacks relate to CDK apps.

Add an approval step (optional)

Add an Approval step between diff and deploy to gate deployment on a manual review of the planned changes.

APPROVAL STEPS HOLD RESOURCES

When you add an Approval step, the underlying machine that runs the pipeline remains active until the approval is resolved. The machine continues to consume compute resources.

From the Pipeline > Execution tab, select + between the diff and deploy steps.
Select Add Step.
Under IACM, select IACM Approval.
Name the approval step and select Apply Changes.
Select Save, then Run your pipeline.

Next steps

This guide introduced you to the core functionalities and setup of Harness IaCM, from creating workspaces to configuring pipelines. To enhance your experience and team efficiency, get the most out of Harness IaCM's built-in reusable features, including:

Workspace Templates: Quickly create new workspaces using pre-defined, customizable templates.
Module Registry: Share and reuse infrastructure modules across teams to standardize deployments.
Default Pipelines: Start with ready-made pipeline templates to accelerate automation and best practices.
Queue Step: Serialize your pipeline execution to track concurrent executions and prevent state conflicts.
Harness CLI: Install the unified CLI to list, get, and execute IaCM workspaces from your terminal.

Go to Harness AI Chat use cases for more use cases on Harness AI, and platform integration.

For an interactive onboarding experience including further use cases and features like **drift detection**, **remote backends**, and **policy enforcement**, check out Harness IaCM self-paced training.

Last updated 1 day ago

Was this helpful?