Skip to main content

The CDK Grain

The AWS CDK grain enables you to orchestrate the deployment and management of AWS resources using the AWS Cloud Development Kit (CDK) as part of a Torque blueprint. This grain allows you to define, provision, and manage AWS infrastructure using familiar programming languages and the CDK toolchain.

info

Currently, the CDK grain supports only the Python SDK.

Overview​

With the AWS CDK grain, you can:

  • Deploy AWS infrastructure as code using CDK apps.
  • Pass environment variables and inputs to parameterize your deployments.
  • Use workspace directories to provide additional files or configuration to your CDK app.
  • Control deployment agents and service accounts for secure execution.

Example: Basic AWS CDK Grain​

spec_version: 2

inputs:
  agent:
    type: agent

grains:
  common:
    kind: aws-cdk
    spec:
      source:
        store: 'cdk-repo'
        path: 'cdk/python/dev/s3'
      agent:
        name: '{{ .inputs.agent }}'

In this example:

  • The CDK app is sourced from the cdk-repo store at cdk/python/dev/s3.
  • The deployment runs on the agent specified by the blueprint input.

Example: Advanced AWS CDK Grain​

spec_version: 2

inputs:
  app_name:
    type: string
  branch:
    type: string
  region:
    type: string
    default: eu-west-1
    allowed-values:
    - eu-west-1
    - us-west-1
  account:
    type: string
  agent:
    type: agent

grains:
  prod-deployment:
    kind: 'aws-cdk'
    spec:
      source:
        store: 'common-cdk'
        path: 'cdk/python/aws/prod'
        branch: '{{ .inputs.branch }}'
      agent:
          name: '{{ .inputs.agent }}'
          service-account: 'prod-sa'
      workspace-directories:
        - source:
            name: 'CONFIG_REPO_PATH'
            store: 'config-repo'
      env-vars:
        - CDK_CONFIG_FILE: '$CONFIG_REPO_PATH/s3'
        - APP: '{{ .inputs.app_name }}'
        - AWS_DEFAULT_REGION: 'eu-west-1'
        - REGION: '{{ .inputs.region }}'
        - ACCOUNT_NAME: '{{ .inputs.account }}'

This example demonstrates:

  • Using a specific branch of a CDK app from a connected store.
  • Passing multiple environment variables, including dynamic values from blueprint inputs.
  • Specifying both the agent and the service account for execution.
  • The environment variable CDK_CONFIG_FILE is set, referencing a workspace directory.
  • The workspace-directories section mounts additional files from another repository.
note

For more details on how to use workspace directories effectively, refer to the Workspace Directories Documentation.

Grain Spec Reference​

source​

Defines where the CDK app is located. Supports:

  • store: The name of the connected repository or asset store.
  • path: Path to the CDK app within the store.
  • branch (optional): Git branch to use.
  • commit (optional): Git commit to use.
  • tag (optional): Git tag to use.

agent​

Specifies the agent on which to run the CDK deployment. You can also specify a service-account for AWS credentials.

env-vars​

A list of environment variables to set for the CDK process. Supports static values and Torque templating.

workspace-directories​

Mounts additional directories from other stores or repositories into the CDK app's workspace. Useful for providing configuration files or dependencies.

Example:

workspace-directories:
  - source:
      store: config-repo
      name: CONFIG_REPO_PATH

inputs​

Inputs can be used to parameterize your CDK deployment, typically referenced in environment variables or within your CDK code.

depends-on​

Defines dependencies on other grains, ensuring this grain runs after the specified ones.

Notes​

  • The AWS CDK grain expects a valid CDK app (e.g., Python, TypeScript, etc.) in the specified path.
  • Ensure that your agent has the necessary AWS credentials and permissions to deploy the resources defined in your CDK app.
  • Use environment variables to pass configuration and secrets securely.
  • You can use Torque's templating engine to inject dynamic values from blueprint inputs or parameters.

See Also​