Deploy stacks in parallel with AWS CDK Pipelines

The AWS CDK Pipelines is a higher-level construct that creates a customized AWS Codepipeline specifically for deploying CDK stacks on AWS Accounts and automatically manages the following:

• Stack dependency order.
• Asset publishing.
• Keeping the pipeline up-to-date as the CDK apps change.
• Using stack outputs later on in the pipeline.

The following tutorial from the AWS docs helps you with setting up an AWS CDK Pipeline. The default CDK Pipeline contains example code to deploy a stack on a single stage.

But if you have multiple test accounts, then it would be nice to add them to the same stage in AWS CDK Pipeline.

To do that, you first have to create a stage and add the first account that you want to deploy an application for:

// Create a stage called Test
const testStage = pipeline.addApplicationStage(
  new PipelineStage(this, 'Test', {
    // Deploy to test account
    env: {
      account: process.env.TEST_AWS_ACCOUNT,
      region: process.env.TEST_AWS_REGION,
    },
    vpcCidr: process.env.TEST_VPC_CIDR,
  })
);

For the deployment of the first application, it does two actions in the stage; “prepare” and “deploy”.

These are linked to “RunOrder: 1” and “RunOrder: 2” respectively. Next, we need to add the second application that we want to add to the “Test” stage.

But before we do we need to update the “RunOrder”, otherwise, the actions that are linked to the second application will be triggered sequentially.

// By subtracting nextSequentialRunOrder with -2 we reset the RunOrder back to 1
testStage.nextSequentialRunOrder(-2);
testStage.addApplication(
  new PipelineTGStage(this, 'Test-tgw', {
    // Deploy to test transit gateway
    env: {
      account: process.env.TEST_TG_AWS_ACCOUNT,
      region: process.env.TEST_TG_AWS_REGION,
    },
  })
);

If we commit the code it will update the CDK Pipelines automatically and the updated stage has two simultaneous stack deploys now: