AWS CodePipeline Actions

---

cdk-constructs: Stable


This package contains Actions that can be used in a CodePipeline.

import aws_cdk.aws_codepipeline as codepipeline
import aws_cdk.aws_codepipeline_actions as codepipeline_actions

Sources

AWS CodeCommit

To use a CodeCommit Repository in a CodePipeline:

repo = codecommit.Repository(self, "Repo",
    repository_name="MyRepo"
)

pipeline = codepipeline.Pipeline(self, "MyPipeline",
    pipeline_name="MyPipeline"
)
source_output = codepipeline.Artifact()
source_action = codepipeline_actions.CodeCommitSourceAction(
    action_name="CodeCommit",
    repository=repo,
    output=source_output
)
pipeline.add_stage(
    stage_name="Source",
    actions=[source_action]
)

If you want to use existing role which can be used by on commit event rule. You can specify the role object in eventRole property.

# repo: codecommit.Repository
event_role = iam.Role.from_role_arn(self, "Event-role", "roleArn")
source_action = codepipeline_actions.CodeCommitSourceAction(
    action_name="CodeCommit",
    repository=repo,
    output=codepipeline.Artifact(),
    event_role=event_role
)

If you want to clone the entire CodeCommit repository (only available for CodeBuild actions), you can set the codeBuildCloneOutput property to true:

# project: codebuild.PipelineProject
# repo: codecommit.Repository

source_output = codepipeline.Artifact()
source_action = codepipeline_actions.CodeCommitSourceAction(
    action_name="CodeCommit",
    repository=repo,
    output=source_output,
    code_build_clone_output=True
)

build_action = codepipeline_actions.CodeBuildAction(
    action_name="CodeBuild",
    project=project,
    input=source_output,  # The build action must use the CodeCommitSourceAction output as input.
    outputs=[codepipeline.Artifact()]
)

The CodeCommit source action emits variables:

# project: codebuild.PipelineProject
# repo: codecommit.Repository

source_output = codepipeline.Artifact()
source_action = codepipeline_actions.CodeCommitSourceAction(
    action_name="CodeCommit",
    repository=repo,
    output=source_output,
    variables_namespace="MyNamespace"
)

# later:

codepipeline_actions.CodeBuildAction(
    action_name="CodeBuild",
    project=project,
    input=source_output,
    environment_variables={
        "COMMIT_ID": codebuild.BuildEnvironmentVariable(
            value=source_action.variables.commit_id
        )
    }
)

GitHub

If you want to use a GitHub repository as the source, you must create:

  • A GitHub Access Token, with scopes repo and admin:repo_hook.

  • A Secrets Manager Secret with the value of the GitHub Access Token. Pick whatever name you want (for example my-github-token). This token can be stored either as Plaintext or as a Secret key/value. If you stored the token as Plaintext, set SecretValue.secretsManager('my-github-token') as the value of oauthToken. If you stored it as a Secret key/value, you must set SecretValue.secretsManager('my-github-token', { jsonField : 'my-github-token' }) as the value of oauthToken.

To use GitHub as the source of a CodePipeline:

# Read the secret from Secrets Manager
pipeline = codepipeline.Pipeline(self, "MyPipeline")
source_output = codepipeline.Artifact()
source_action = codepipeline_actions.GitHubSourceAction(
    action_name="GitHub_Source",
    owner="awslabs",
    repo="aws-cdk",
    oauth_token=SecretValue.secrets_manager("my-github-token"),
    output=source_output,
    branch="develop"
)
pipeline.add_stage(
    stage_name="Source",
    actions=[source_action]
)

The GitHub source action emits variables:

# source_output: codepipeline.Artifact
# project: codebuild.PipelineProject


source_action = codepipeline_actions.GitHubSourceAction(
    action_name="Github_Source",
    output=source_output,
    owner="my-owner",
    repo="my-repo",
    oauth_token=SecretValue.secrets_manager("my-github-token"),
    variables_namespace="MyNamespace"
)

# later:

codepipeline_actions.CodeBuildAction(
    action_name="CodeBuild",
    project=project,
    input=source_output,
    environment_variables={
        "COMMIT_URL": codebuild.BuildEnvironmentVariable(
            value=source_action.variables.commit_url
        )
    }
)

BitBucket

CodePipeline can use a BitBucket Git repository as a source:

Note: you have to manually connect CodePipeline through the AWS Console with your BitBucket account. This is a one-time operation for a given AWS account in a given region. The simplest way to do that is to either start creating a new CodePipeline, or edit an existing one, while being logged in to BitBucket. Choose BitBucket as the source, and grant CodePipeline permissions to your BitBucket account. Copy & paste the Connection ARN that you get in the console, or use the codestar-connections list-connections AWS CLI operation to find it. After that, you can safely abort creating or editing the pipeline - the connection has already been created.

source_output = codepipeline.Artifact()
source_action = codepipeline_actions.CodeStarConnectionsSourceAction(
    action_name="BitBucket_Source",
    owner="aws",
    repo="aws-cdk",
    output=source_output,
    connection_arn="arn:aws:codestar-connections:us-east-1:123456789012:connection/12345678-abcd-12ab-34cdef5678gh"
)

You can also use the CodeStarConnectionsSourceAction to connect to GitHub, in the same way (you just have to select GitHub as the source when creating the connection in the console).

Similarly to GitHubSourceAction, CodeStarConnectionsSourceAction also emits the variables:

# project: codebuild.Project


source_output = codepipeline.Artifact()
source_action = codepipeline_actions.CodeStarConnectionsSourceAction(
    action_name="BitBucket_Source",
    owner="aws",
    repo="aws-cdk",
    output=source_output,
    connection_arn="arn:aws:codestar-connections:us-east-1:123456789012:connection/12345678-abcd-12ab-34cdef5678gh",
    variables_namespace="SomeSpace"
)

# later:

codepipeline_actions.CodeBuildAction(
    action_name="CodeBuild",
    project=project,
    input=source_output,
    environment_variables={
        "COMMIT_ID": codebuild.BuildEnvironmentVariable(
            value=source_action.variables.commit_id
        )
    }
)

AWS S3 Source

To use an S3 Bucket as a source in CodePipeline:

source_bucket = s3.Bucket(self, "MyBucket",
    versioned=True
)

pipeline = codepipeline.Pipeline(self, "MyPipeline")
source_output = codepipeline.Artifact()
source_action = codepipeline_actions.S3SourceAction(
    action_name="S3Source",
    bucket=source_bucket,
    bucket_key="path/to/file.zip",
    output=source_output
)
pipeline.add_stage(
    stage_name="Source",
    actions=[source_action]
)

The region of the action will be determined by the region the bucket itself is in. When using a newly created bucket, that region will be taken from the stack the bucket belongs to; for an imported bucket, you can specify the region explicitly:

source_bucket = s3.Bucket.from_bucket_attributes(self, "SourceBucket",
    bucket_name="my-bucket",
    region="ap-southeast-1"
)

By default, the Pipeline will poll the Bucket to detect changes. You can change that behavior to use CloudWatch Events by setting the trigger property to S3Trigger.EVENTS (it’s S3Trigger.POLL by default). If you do that, make sure the source Bucket is part of an AWS CloudTrail Trail - otherwise, the CloudWatch Events will not be emitted, and your Pipeline will not react to changes in the Bucket. You can do it through the CDK:

import aws_cdk.aws_cloudtrail as cloudtrail

# source_bucket: s3.Bucket

source_output = codepipeline.Artifact()
key = "some/key.zip"
trail = cloudtrail.Trail(self, "CloudTrail")
trail.add_s3_event_selector([cloudtrail.S3EventSelector(
    bucket=source_bucket,
    object_prefix=key
)],
    read_write_type=cloudtrail.ReadWriteType.WRITE_ONLY
)
source_action = codepipeline_actions.S3SourceAction(
    action_name="S3Source",
    bucket_key=key,
    bucket=source_bucket,
    output=source_output,
    trigger=codepipeline_actions.S3Trigger.EVENTS
)

The S3 source action emits variables:

# source_bucket: s3.Bucket

# later:
# project: codebuild.PipelineProject
key = "some/key.zip"
source_output = codepipeline.Artifact()
source_action = codepipeline_actions.S3SourceAction(
    action_name="S3Source",
    bucket_key=key,
    bucket=source_bucket,
    output=source_output,
    variables_namespace="MyNamespace"
)
codepipeline_actions.CodeBuildAction(
    action_name="CodeBuild",
    project=project,
    input=source_output,
    environment_variables={
        "VERSION_ID": codebuild.BuildEnvironmentVariable(
            value=source_action.variables.version_id
        )
    }
)

AWS ECR

To use an ECR Repository as a source in a Pipeline:

import aws_cdk.aws_ecr as ecr

# ecr_repository: ecr.Repository

pipeline = codepipeline.Pipeline(self, "MyPipeline")
source_output = codepipeline.Artifact()
source_action = codepipeline_actions.EcrSourceAction(
    action_name="ECR",
    repository=ecr_repository,
    image_tag="some-tag",  # optional, default: 'latest'
    output=source_output
)
pipeline.add_stage(
    stage_name="Source",
    actions=[source_action]
)

The ECR source action emits variables:

import aws_cdk.aws_ecr as ecr
# ecr_repository: ecr.Repository

# later:
# project: codebuild.PipelineProject


source_output = codepipeline.Artifact()
source_action = codepipeline_actions.EcrSourceAction(
    action_name="Source",
    output=source_output,
    repository=ecr_repository,
    variables_namespace="MyNamespace"
)
codepipeline_actions.CodeBuildAction(
    action_name="CodeBuild",
    project=project,
    input=source_output,
    environment_variables={
        "IMAGE_URI": codebuild.BuildEnvironmentVariable(
            value=source_action.variables.image_uri
        )
    }
)

Build & test

AWS CodeBuild

Example of a CodeBuild Project used in a Pipeline, alongside CodeCommit:

# project: codebuild.PipelineProject

repository = codecommit.Repository(self, "MyRepository",
    repository_name="MyRepository"
)
project = codebuild.PipelineProject(self, "MyProject")

source_output = codepipeline.Artifact()
source_action = codepipeline_actions.CodeCommitSourceAction(
    action_name="CodeCommit",
    repository=repository,
    output=source_output
)
build_action = codepipeline_actions.CodeBuildAction(
    action_name="CodeBuild",
    project=project,
    input=source_output,
    outputs=[codepipeline.Artifact()],  # optional
    execute_batch_build=True,  # optional, defaults to false
    combine_batch_build_artifacts=True
)

codepipeline.Pipeline(self, "MyPipeline",
    stages=[codepipeline.StageProps(
        stage_name="Source",
        actions=[source_action]
    ), codepipeline.StageProps(
        stage_name="Build",
        actions=[build_action]
    )
    ]
)

The default category of the CodeBuild Action is Build; if you want a Test Action instead, override the type property:

# project: codebuild.PipelineProject

source_output = codepipeline.Artifact()
test_action = codepipeline_actions.CodeBuildAction(
    action_name="IntegrationTest",
    project=project,
    input=source_output,
    type=codepipeline_actions.CodeBuildActionType.TEST
)

Multiple inputs and outputs

When you want to have multiple inputs and/or outputs for a Project used in a Pipeline, instead of using the secondarySources and secondaryArtifacts properties of the Project class, you need to use the extraInputs and outputs properties of the CodeBuild CodePipeline Actions. Example:

# repository1: codecommit.Repository
# repository2: codecommit.Repository

# project: codebuild.PipelineProject

source_output1 = codepipeline.Artifact()
source_action1 = codepipeline_actions.CodeCommitSourceAction(
    action_name="Source1",
    repository=repository1,
    output=source_output1
)
source_output2 = codepipeline.Artifact("source2")
source_action2 = codepipeline_actions.CodeCommitSourceAction(
    action_name="Source2",
    repository=repository2,
    output=source_output2
)
build_action = codepipeline_actions.CodeBuildAction(
    action_name="Build",
    project=project,
    input=source_output1,
    extra_inputs=[source_output2
    ],
    outputs=[
        codepipeline.Artifact("artifact1"),  # for better buildspec readability - see below
        codepipeline.Artifact("artifact2")
    ]
)

Note: when a CodeBuild Action in a Pipeline has more than one output, it only uses the secondary-artifacts field of the buildspec, never the primary output specification directly under artifacts. Because of that, it pays to explicitly name all output artifacts of that Action, like we did above, so that you know what name to use in the buildspec.

Example buildspec for the above project:

project = codebuild.PipelineProject(self, "MyProject",
    build_spec=codebuild.BuildSpec.from_object({
        "version": "0.2",
        "phases": {
            "build": {
                "commands": []
            }
        },
        "artifacts": {
            "secondary-artifacts": {
                "artifact1": {},
                "artifact2": {}
            }
        }
    })
)

Variables

The CodeBuild action emits variables. Unlike many other actions, the variables are not static, but dynamic, defined in the buildspec, in the ‘exported-variables’ subsection of the ‘env’ section. Example:

# later:
# project: codebuild.PipelineProject
source_output = codepipeline.Artifact()
build_action = codepipeline_actions.CodeBuildAction(
    action_name="Build1",
    input=source_output,
    project=codebuild.PipelineProject(self, "Project",
        build_spec=codebuild.BuildSpec.from_object({
            "version": "0.2",
            "env": {
                "exported-variables": ["MY_VAR"
                ]
            },
            "phases": {
                "build": {
                    "commands": "export MY_VAR="some value""
                }
            }
        })
    ),
    variables_namespace="MyNamespace"
)
codepipeline_actions.CodeBuildAction(
    action_name="CodeBuild",
    project=project,
    input=source_output,
    environment_variables={
        "MyVar": codebuild.BuildEnvironmentVariable(
            value=build_action.variable("MY_VAR")
        )
    }
)

Jenkins

In order to use Jenkins Actions in the Pipeline, you first need to create a JenkinsProvider:

jenkins_provider = codepipeline_actions.JenkinsProvider(self, "JenkinsProvider",
    provider_name="MyJenkinsProvider",
    server_url="http://my-jenkins.com:8080",
    version="2"
)

If you’ve registered a Jenkins provider in a different CDK app, or outside the CDK (in the CodePipeline AWS Console, for example), you can import it:

jenkins_provider = codepipeline_actions.JenkinsProvider.from_jenkins_provider_attributes(self, "JenkinsProvider",
    provider_name="MyJenkinsProvider",
    server_url="http://my-jenkins.com:8080",
    version="2"
)

Note that a Jenkins provider (identified by the provider name-category(build/test)-version tuple) must always be registered in the given account, in the given AWS region, before it can be used in CodePipeline.

With a JenkinsProvider, we can create a Jenkins Action:

# jenkins_provider: codepipeline_actions.JenkinsProvider

build_action = codepipeline_actions.JenkinsAction(
    action_name="JenkinsBuild",
    jenkins_provider=jenkins_provider,
    project_name="MyProject",
    type=codepipeline_actions.JenkinsActionType.BUILD
)

Deploy

AWS CloudFormation

This module contains Actions that allows you to deploy to CloudFormation from AWS CodePipeline.

For example, the following code fragment defines a pipeline that automatically deploys a CloudFormation template directly from a CodeCommit repository, with a manual approval step in between to confirm the changes:

# Source stage: read from repository
repo = codecommit.Repository(stack, "TemplateRepo",
    repository_name="template-repo"
)
source_output = codepipeline.Artifact("SourceArtifact")
source = cpactions.CodeCommitSourceAction(
    action_name="Source",
    repository=repo,
    output=source_output,
    trigger=cpactions.CodeCommitTrigger.POLL
)
source_stage = {
    "stage_name": "Source",
    "actions": [source]
}

# Deployment stage: create and deploy changeset with manual approval
stack_name = "OurStack"
change_set_name = "StagedChangeSet"

prod_stage = {
    "stage_name": "Deploy",
    "actions": [
        cpactions.CloudFormationCreateReplaceChangeSetAction(
            action_name="PrepareChanges",
            stack_name=stack_name,
            change_set_name=change_set_name,
            admin_permissions=True,
            template_path=source_output.at_path("template.yaml"),
            run_order=1
        ),
        cpactions.ManualApprovalAction(
            action_name="ApproveChanges",
            run_order=2
        ),
        cpactions.CloudFormationExecuteChangeSetAction(
            action_name="ExecuteChanges",
            stack_name=stack_name,
            change_set_name=change_set_name,
            run_order=3
        )
    ]
}

codepipeline.Pipeline(stack, "Pipeline",
    stages=[source_stage, prod_stage
    ]
)

See the AWS documentation for more details about using CloudFormation in CodePipeline.

Actions for updating individual CloudFormation Stacks

This package contains the following CloudFormation actions:

  • CloudFormationCreateUpdateStackAction - Deploy a CloudFormation template directly from the pipeline. The indicated stack is created, or updated if it already exists. If the stack is in a failure state, deployment will fail (unless replaceOnFailure is set to true, in which case it will be destroyed and recreated).

  • CloudFormationDeleteStackAction - Delete the stack with the given name.

  • CloudFormationCreateReplaceChangeSetAction - Prepare a change set to be applied later. You will typically use change sets if you want to manually verify the changes that are being staged, or if you want to separate the people (or system) preparing the changes from the people (or system) applying the changes.

  • CloudFormationExecuteChangeSetAction - Execute a change set prepared previously.

Actions for deploying CloudFormation StackSets to multiple accounts

You can use CloudFormation StackSets to deploy the same CloudFormation template to multiple accounts in a managed way. If you use AWS Organizations, StackSets can be deployed to all accounts in a particular Organizational Unit (OU), and even automatically to new accounts as soon as they are added to a particular OU. For more information, see the Working with StackSets section of the CloudFormation developer guide.

The actions available for updating StackSets are:

  • CloudFormationDeployStackSetAction - Create or update a CloudFormation StackSet directly from the pipeline, optionally immediately create and update Stack Instances as well.

  • CloudFormationDeployStackInstancesAction - Update outdated Stack Instaces using the current version of the StackSet.

Here’s an example of using both of these actions:

# pipeline: codepipeline.Pipeline
# source_output: codepipeline.Artifact


pipeline.add_stage(
    stage_name="DeployStackSets",
    actions=[
        # First, update the StackSet itself with the newest template
        codepipeline_actions.CloudFormationDeployStackSetAction(
            action_name="UpdateStackSet",
            run_order=1,
            stack_set_name="MyStackSet",
            template=codepipeline_actions.StackSetTemplate.from_artifact_path(source_output.at_path("template.yaml")),

            # Change this to 'StackSetDeploymentModel.organizations()' if you want to deploy to OUs
            deployment_model=codepipeline_actions.StackSetDeploymentModel.self_managed(),
            # This deploys to a set of accounts
            stack_instances=codepipeline_actions.StackInstances.in_accounts(["111111111111"], ["us-east-1", "eu-west-1"])
        ),

        # Afterwards, update/create additional instances in other accounts
        codepipeline_actions.CloudFormationDeployStackInstancesAction(
            action_name="AddMoreInstances",
            run_order=2,
            stack_set_name="MyStackSet",
            stack_instances=codepipeline_actions.StackInstances.in_accounts(["222222222222", "333333333333"], ["us-east-1", "eu-west-1"])
        )
    ]
)

Lambda deployed through CodePipeline

If you want to deploy your Lambda through CodePipeline, and you don’t use assets (for example, because your CDK code and Lambda code are separate), you can use a special Lambda Code class, CfnParametersCode. Note that your Lambda must be in a different Stack than your Pipeline. The Lambda itself will be deployed, alongside the entire Stack it belongs to, using a CloudFormation CodePipeline Action. Example:

lambda_stack = cdk.Stack(app, "LambdaStack")
lambda_code = lambda_.Code.from_cfn_parameters()
lambda_.Function(lambda_stack, "Lambda",
    code=lambda_code,
    handler="index.handler",
    runtime=lambda_.Runtime.NODEJS_14_X
)
# other resources that your Lambda needs, added to the lambdaStack...

pipeline_stack = cdk.Stack(app, "PipelineStack")
pipeline = codepipeline.Pipeline(pipeline_stack, "Pipeline")

# add the source code repository containing this code to your Pipeline,
# and the source code of the Lambda Function, if they're separate
cdk_source_output = codepipeline.Artifact()
cdk_source_action = codepipeline_actions.CodeCommitSourceAction(
    repository=codecommit.Repository(pipeline_stack, "CdkCodeRepo",
        repository_name="CdkCodeRepo"
    ),
    action_name="CdkCode_Source",
    output=cdk_source_output
)
lambda_source_output = codepipeline.Artifact()
lambda_source_action = codepipeline_actions.CodeCommitSourceAction(
    repository=codecommit.Repository(pipeline_stack, "LambdaCodeRepo",
        repository_name="LambdaCodeRepo"
    ),
    action_name="LambdaCode_Source",
    output=lambda_source_output
)
pipeline.add_stage(
    stage_name="Source",
    actions=[cdk_source_action, lambda_source_action]
)

# synthesize the Lambda CDK template, using CodeBuild
# the below values are just examples, assuming your CDK code is in TypeScript/JavaScript -
# adjust the build environment and/or commands accordingly
cdk_build_project = codebuild.Project(pipeline_stack, "CdkBuildProject",
    environment=codebuild.BuildEnvironment(
        build_image=codebuild.LinuxBuildImage.UBUNTU_14_04_NODEJS_10_1_0
    ),
    build_spec=codebuild.BuildSpec.from_object({
        "version": "0.2",
        "phases": {
            "install": {
                "commands": "npm install"
            },
            "build": {
                "commands": ["npm run build", "npm run cdk synth LambdaStack -- -o ."
                ]
            }
        },
        "artifacts": {
            "files": "LambdaStack.template.yaml"
        }
    })
)
cdk_build_output = codepipeline.Artifact()
cdk_build_action = codepipeline_actions.CodeBuildAction(
    action_name="CDK_Build",
    project=cdk_build_project,
    input=cdk_source_output,
    outputs=[cdk_build_output]
)

# build your Lambda code, using CodeBuild
# again, this example assumes your Lambda is written in TypeScript/JavaScript -
# make sure to adjust the build environment and/or commands if they don't match your specific situation
lambda_build_project = codebuild.Project(pipeline_stack, "LambdaBuildProject",
    environment=codebuild.BuildEnvironment(
        build_image=codebuild.LinuxBuildImage.UBUNTU_14_04_NODEJS_10_1_0
    ),
    build_spec=codebuild.BuildSpec.from_object({
        "version": "0.2",
        "phases": {
            "install": {
                "commands": "npm install"
            },
            "build": {
                "commands": "npm run build"
            }
        },
        "artifacts": {
            "files": ["index.js", "node_modules/**/*"
            ]
        }
    })
)
lambda_build_output = codepipeline.Artifact()
lambda_build_action = codepipeline_actions.CodeBuildAction(
    action_name="Lambda_Build",
    project=lambda_build_project,
    input=lambda_source_output,
    outputs=[lambda_build_output]
)

pipeline.add_stage(
    stage_name="Build",
    actions=[cdk_build_action, lambda_build_action]
)

# finally, deploy your Lambda Stack
pipeline.add_stage(
    stage_name="Deploy",
    actions=[
        codepipeline_actions.CloudFormationCreateUpdateStackAction(
            action_name="Lambda_CFN_Deploy",
            template_path=cdk_build_output.at_path("LambdaStack.template.yaml"),
            stack_name="LambdaStackDeployedName",
            admin_permissions=True,
            parameter_overrides=lambda_code.assign(lambda_build_output.s3_location),
            extra_inputs=[lambda_build_output
            ]
        )
    ]
)

Cross-account actions

If you want to update stacks in a different account, pass the account property when creating the action:

source_output = codepipeline.Artifact()
codepipeline_actions.CloudFormationCreateUpdateStackAction(
    action_name="CloudFormationCreateUpdate",
    stack_name="MyStackName",
    admin_permissions=True,
    template_path=source_output.at_path("template.yaml"),
    account="123456789012"
)

This will create a new stack, called <PipelineStackName>-support-123456789012, in your App, that will contain the role that the pipeline will assume in account 123456789012 before executing this action. This support stack will automatically be deployed before the stack containing the pipeline.

You can also pass a role explicitly when creating the action - in that case, the account property is ignored, and the action will operate in the same account the role belongs to:

from aws_cdk.core import PhysicalName

# in stack for account 123456789012...
# other_account_stack: Stack

action_role = iam.Role(other_account_stack, "ActionRole",
    assumed_by=iam.AccountPrincipal("123456789012"),
    # the role has to have a physical name set
    role_name=PhysicalName.GENERATE_IF_NEEDED
)

# in the pipeline stack...
source_output = codepipeline.Artifact()
codepipeline_actions.CloudFormationCreateUpdateStackAction(
    action_name="CloudFormationCreateUpdate",
    stack_name="MyStackName",
    admin_permissions=True,
    template_path=source_output.at_path("template.yaml"),
    role=action_role
)

AWS CodeDeploy

Server deployments

To use CodeDeploy for EC2/on-premise deployments in a Pipeline:

# deployment_group: codedeploy.ServerDeploymentGroup
pipeline = codepipeline.Pipeline(self, "MyPipeline",
    pipeline_name="MyPipeline"
)

# add the source and build Stages to the Pipeline...
build_output = codepipeline.Artifact()
deploy_action = codepipeline_actions.CodeDeployServerDeployAction(
    action_name="CodeDeploy",
    input=build_output,
    deployment_group=deployment_group
)
pipeline.add_stage(
    stage_name="Deploy",
    actions=[deploy_action]
)
Lambda deployments

To use CodeDeploy for blue-green Lambda deployments in a Pipeline:

lambda_code = lambda_.Code.from_cfn_parameters()
func = lambda_.Function(self, "Lambda",
    code=lambda_code,
    handler="index.handler",
    runtime=lambda_.Runtime.NODEJS_12_X
)
# used to make sure each CDK synthesis produces a different Version
version = func.current_version
alias = lambda_.Alias(self, "LambdaAlias",
    alias_name="Prod",
    version=version
)

codedeploy.LambdaDeploymentGroup(self, "DeploymentGroup",
    alias=alias,
    deployment_config=codedeploy.LambdaDeploymentConfig.LINEAR_10PERCENT_EVERY_1MINUTE
)

Then, you need to create your Pipeline Stack, where you will define your Pipeline, and deploy the lambdaStack using a CloudFormation CodePipeline Action (see above for a complete example).

ECS

CodePipeline can deploy an ECS service. The deploy Action receives one input Artifact which contains the image definition file:

import aws_cdk.aws_ecs as ecs

# service: ecs.FargateService

pipeline = codepipeline.Pipeline(self, "MyPipeline")
build_output = codepipeline.Artifact()
deploy_stage = pipeline.add_stage(
    stage_name="Deploy",
    actions=[
        codepipeline_actions.EcsDeployAction(
            action_name="DeployAction",
            service=service,
            # if your file is called imagedefinitions.json,
            # use the `input` property,
            # and leave out the `imageFile` property
            input=build_output,
            # if your file name is _not_ imagedefinitions.json,
            # use the `imageFile` property,
            # and leave out the `input` property
            image_file=build_output.at_path("imageDef.json"),
            deployment_timeout=Duration.minutes(60)
        )
    ]
)

Deploying ECS applications to existing services

CodePipeline can deploy to an existing ECS service which uses the ECS service ARN format that contains the Cluster name. This also works if the service is in a different account and/or region than the pipeline:

import aws_cdk.aws_ecs as ecs


service = ecs.BaseService.from_service_arn_with_cluster(self, "EcsService", "arn:aws:ecs:us-east-1:123456789012:service/myClusterName/myServiceName")
pipeline = codepipeline.Pipeline(self, "MyPipeline")
build_output = codepipeline.Artifact()
# add source and build stages to the pipeline as usual...
deploy_stage = pipeline.add_stage(
    stage_name="Deploy",
    actions=[
        codepipeline_actions.EcsDeployAction(
            action_name="DeployAction",
            service=service,
            input=build_output
        )
    ]
)

When deploying across accounts, especially in a CDK Pipelines self-mutating pipeline, it is recommended to provide the role property to the EcsDeployAction. The Role will need to have permissions assigned to it for ECS deployment. See the CodePipeline documentation for the permissions needed.

Deploying ECS applications stored in a separate source code repository

The idiomatic CDK way of deploying an ECS application is to have your Dockerfiles and your CDK code in the same source code repository, leveraging Docker Assets, and use the CDK Pipelines module.

However, if you want to deploy a Docker application whose source code is kept in a separate version control repository than the CDK code, you can use the TagParameterContainerImage class from the ECS module. Here’s an example:

#
# This is the Stack containing a simple ECS Service that uses the provided ContainerImage.
#
class EcsAppStack(cdk.Stack):
    def __init__(self, scope, id, *, image, description=None, env=None, stackName=None, tags=None, synthesizer=None, terminationProtection=None, analyticsReporting=None):
        super().__init__(scope, id, image=image, description=description, env=env, stackName=stackName, tags=tags, synthesizer=synthesizer, terminationProtection=terminationProtection, analyticsReporting=analyticsReporting)

        task_definition = ecs.TaskDefinition(self, "TaskDefinition",
            compatibility=ecs.Compatibility.FARGATE,
            cpu="1024",
            memory_mi_b="2048"
        )
        task_definition.add_container("AppContainer",
            image=image
        )
        ecs.FargateService(self, "EcsService",
            task_definition=task_definition,
            cluster=ecs.Cluster(self, "Cluster",
                vpc=ec2.Vpc(self, "Vpc",
                    max_azs=1
                )
            )
        )

#
# This is the Stack containing the CodePipeline definition that deploys an ECS Service.
#
class PipelineStack(cdk.Stack):

    def __init__(self, scope, id, *, description=None, env=None, stackName=None, tags=None, synthesizer=None, terminationProtection=None, analyticsReporting=None):
        super().__init__(scope, id, description=description, env=env, stackName=stackName, tags=tags, synthesizer=synthesizer, terminationProtection=terminationProtection, analyticsReporting=analyticsReporting)

        # ********* ECS part ****************

        # this is the ECR repository where the built Docker image will be pushed
        app_ecr_repo = ecr.Repository(self, "EcsDeployRepository")
        # the build that creates the Docker image, and pushes it to the ECR repo
        app_code_docker_build = codebuild.PipelineProject(self, "AppCodeDockerImageBuildAndPushProject",
            environment=codebuild.BuildEnvironment(
                # we need to run Docker
                privileged=True
            ),
            build_spec=codebuild.BuildSpec.from_object({
                "version": "0.2",
                "phases": {
                    "build": {
                        "commands": ["$(aws ecr get-login --region $AWS_DEFAULT_REGION --no-include-email)", "docker build -t $REPOSITORY_URI:$CODEBUILD_RESOLVED_SOURCE_VERSION ."
                        ]
                    },
                    "post_build": {
                        "commands": ["docker push $REPOSITORY_URI:$CODEBUILD_RESOLVED_SOURCE_VERSION", "export imageTag=$CODEBUILD_RESOLVED_SOURCE_VERSION"
                        ]
                    }
                },
                "env": {
                    # save the imageTag environment variable as a CodePipeline Variable
                    "exported-variables": ["imageTag"
                    ]
                }
            }),
            environment_variables={
                "REPOSITORY_URI": codebuild.BuildEnvironmentVariable(
                    value=app_ecr_repo.repository_uri
                )
            }
        )
        # needed for `docker push`
        app_ecr_repo.grant_pull_push(app_code_docker_build)
        # create the ContainerImage used for the ECS application Stack
        self.tag_parameter_container_image = ecs.TagParameterContainerImage(app_ecr_repo)

        cdk_code_build = codebuild.PipelineProject(self, "CdkCodeBuildProject",
            build_spec=codebuild.BuildSpec.from_object({
                "version": "0.2",
                "phases": {
                    "install": {
                        "commands": ["npm install"
                        ]
                    },
                    "build": {
                        "commands": ["npx cdk synth --verbose"
                        ]
                    }
                },
                "artifacts": {
                    # store the entire Cloud Assembly as the output artifact
                    "base-directory": "cdk.out",
                    "files": "**/*"
                }
            })
        )

        # ********* Pipeline part ****************

        app_code_source_output = codepipeline.Artifact()
        cdk_code_source_output = codepipeline.Artifact()
        cdk_code_build_output = codepipeline.Artifact()
        app_code_build_action = codepipeline_actions.CodeBuildAction(
            action_name="AppCodeDockerImageBuildAndPush",
            project=app_code_docker_build,
            input=app_code_source_output
        )
        codepipeline.Pipeline(self, "CodePipelineDeployingEcsApplication",
            artifact_bucket=s3.Bucket(self, "ArtifactBucket",
                removal_policy=cdk.RemovalPolicy.DESTROY
            ),
            stages=[codepipeline.StageProps(
                stage_name="Source",
                actions=[
                    # this is the Action that takes the source of your application code
                    codepipeline_actions.CodeCommitSourceAction(
                        action_name="AppCodeSource",
                        repository=codecommit.Repository(self, "AppCodeSourceRepository", repository_name="AppCodeSourceRepository"),
                        output=app_code_source_output
                    ),
                    # this is the Action that takes the source of your CDK code
                    # (which would probably include this Pipeline code as well)
                    codepipeline_actions.CodeCommitSourceAction(
                        action_name="CdkCodeSource",
                        repository=codecommit.Repository(self, "CdkCodeSourceRepository", repository_name="CdkCodeSourceRepository"),
                        output=cdk_code_source_output
                    )
                ]
            ), codepipeline.StageProps(
                stage_name="Build",
                actions=[app_code_build_action,
                    codepipeline_actions.CodeBuildAction(
                        action_name="CdkCodeBuildAndSynth",
                        project=cdk_code_build,
                        input=cdk_code_source_output,
                        outputs=[cdk_code_build_output]
                    )
                ]
            ), codepipeline.StageProps(
                stage_name="Deploy",
                actions=[
                    codepipeline_actions.CloudFormationCreateUpdateStackAction(
                        action_name="CFN_Deploy",
                        stack_name="SampleEcsStackDeployedFromCodePipeline",
                        # this name has to be the same name as used below in the CDK code for the application Stack
                        template_path=cdk_code_build_output.at_path("EcsStackDeployedInPipeline.template.json"),
                        admin_permissions=True,
                        parameter_overrides={
                            # read the tag pushed to the ECR repository from the CodePipeline Variable saved by the application build step,
                            # and pass it as the CloudFormation Parameter for the tag
                            "self.tag_parameter_container_image.tag_parameter_name": app_code_build_action.variable("imageTag")
                        }
                    )
                ]
            )
            ]
        )

app = cdk.App()

# the CodePipeline Stack needs to be created first
pipeline_stack = PipelineStack(app, "aws-cdk-pipeline-ecs-separate-sources")
# we supply the image to the ECS application Stack from the CodePipeline Stack
EcsAppStack(app, "EcsStackDeployedInPipeline",
    image=pipeline_stack.tag_parameter_container_image
)

AWS S3 Deployment

To use an S3 Bucket as a deployment target in CodePipeline:

source_output = codepipeline.Artifact()
target_bucket = s3.Bucket(self, "MyBucket")

pipeline = codepipeline.Pipeline(self, "MyPipeline")
deploy_action = codepipeline_actions.S3DeployAction(
    action_name="S3Deploy",
    bucket=target_bucket,
    input=source_output
)
deploy_stage = pipeline.add_stage(
    stage_name="Deploy",
    actions=[deploy_action]
)

Invalidating the CloudFront cache when deploying to S3

There is currently no native support in CodePipeline for invalidating a CloudFront cache after deployment. One workaround is to add another build step after the deploy step, and use the AWS CLI to invalidate the cache:

# Create a Cloudfront Web Distribution
import aws_cdk.aws_cloudfront as cloudfront
# distribution: cloudfront.Distribution


# Create the build project that will invalidate the cache
invalidate_build_project = codebuild.PipelineProject(self, "InvalidateProject",
    build_spec=codebuild.BuildSpec.from_object({
        "version": "0.2",
        "phases": {
            "build": {
                "commands": ["aws cloudfront create-invalidation --distribution-id ${CLOUDFRONT_ID} --paths "/*""
                ]
            }
        }
    }),
    environment_variables={
        "CLOUDFRONT_ID": codebuild.BuildEnvironmentVariable(value=distribution.distribution_id)
    }
)

# Add Cloudfront invalidation permissions to the project
distribution_arn = f"arn:aws:cloudfront::{this.account}:distribution/{distribution.distributionId}"
invalidate_build_project.add_to_role_policy(iam.PolicyStatement(
    resources=[distribution_arn],
    actions=["cloudfront:CreateInvalidation"
    ]
))

# Create the pipeline (here only the S3 deploy and Invalidate cache build)
deploy_bucket = s3.Bucket(self, "DeployBucket")
deploy_input = codepipeline.Artifact()
codepipeline.Pipeline(self, "Pipeline",
    stages=[codepipeline.StageProps(
        stage_name="Deploy",
        actions=[
            codepipeline_actions.S3DeployAction(
                action_name="S3Deploy",
                bucket=deploy_bucket,
                input=deploy_input,
                run_order=1
            ),
            codepipeline_actions.CodeBuildAction(
                action_name="InvalidateCache",
                project=invalidate_build_project,
                input=deploy_input,
                run_order=2
            )
        ]
    )
    ]
)

Alexa Skill

You can deploy to Alexa using CodePipeline with the following Action:

# Read the secrets from ParameterStore
client_id = SecretValue.secrets_manager("AlexaClientId")
client_secret = SecretValue.secrets_manager("AlexaClientSecret")
refresh_token = SecretValue.secrets_manager("AlexaRefreshToken")

# Add deploy action
source_output = codepipeline.Artifact()
codepipeline_actions.AlexaSkillDeployAction(
    action_name="DeploySkill",
    run_order=1,
    input=source_output,
    client_id=client_id.to_string(),
    client_secret=client_secret,
    refresh_token=refresh_token,
    skill_id="amzn1.ask.skill.12345678-1234-1234-1234-123456789012"
)

If you need manifest overrides you can specify them as parameterOverridesArtifact in the action:

# Deploy some CFN change set and store output
execute_output = codepipeline.Artifact("CloudFormation")
execute_change_set_action = codepipeline_actions.CloudFormationExecuteChangeSetAction(
    action_name="ExecuteChangesTest",
    run_order=2,
    stack_name="MyStack",
    change_set_name="MyChangeSet",
    output_file_name="overrides.json",
    output=execute_output
)

# Provide CFN output as manifest overrides
client_id = SecretValue.secrets_manager("AlexaClientId")
client_secret = SecretValue.secrets_manager("AlexaClientSecret")
refresh_token = SecretValue.secrets_manager("AlexaRefreshToken")
source_output = codepipeline.Artifact()
codepipeline_actions.AlexaSkillDeployAction(
    action_name="DeploySkill",
    run_order=1,
    input=source_output,
    parameter_overrides_artifact=execute_output,
    client_id=client_id.to_string(),
    client_secret=client_secret,
    refresh_token=refresh_token,
    skill_id="amzn1.ask.skill.12345678-1234-1234-1234-123456789012"
)

AWS Service Catalog

You can deploy a CloudFormation template to an existing Service Catalog product with the following Action:

cdk_build_output = codepipeline.Artifact()
service_catalog_deploy_action = codepipeline_actions.ServiceCatalogDeployActionBeta1(
    action_name="ServiceCatalogDeploy",
    template_path=cdk_build_output.at_path("Sample.template.json"),
    product_version_name="Version - " + Date.now.to_string,
    product_version_description="This is a version from the pipeline with a new description.",
    product_id="prod-XXXXXXXX"
)

Approve & invoke

Manual approval Action

This package contains an Action that stops the Pipeline until someone manually clicks the approve button:

import aws_cdk.aws_sns as sns


pipeline = codepipeline.Pipeline(self, "MyPipeline")
approve_stage = pipeline.add_stage(stage_name="Approve")
manual_approval_action = codepipeline_actions.ManualApprovalAction(
    action_name="Approve",
    notification_topic=sns.Topic(self, "Topic"),  # optional
    notify_emails=["some_email@example.com"
    ],  # optional
    additional_information="additional info"
)
approve_stage.add_action(manual_approval_action)

If the notificationTopic has not been provided, but notifyEmails were, a new SNS Topic will be created (and accessible through the notificationTopic property of the Action).

If you want to grant a principal permissions to approve the changes, you can invoke the method grantManualApproval passing it a IGrantable:

pipeline = codepipeline.Pipeline(self, "MyPipeline")
approve_stage = pipeline.add_stage(stage_name="Approve")
manual_approval_action = codepipeline_actions.ManualApprovalAction(
    action_name="Approve"
)
approve_stage.add_action(manual_approval_action)

role = iam.Role.from_role_arn(self, "Admin", Arn.format(ArnComponents(service="iam", resource="role", resource_name="Admin"), self))
manual_approval_action.grant_manual_approval(role)

AWS Lambda

This module contains an Action that allows you to invoke a Lambda function in a Pipeline:

# fn: lambda.Function

pipeline = codepipeline.Pipeline(self, "MyPipeline")
lambda_action = codepipeline_actions.LambdaInvokeAction(
    action_name="Lambda",
    lambda_=fn
)
pipeline.add_stage(
    stage_name="Lambda",
    actions=[lambda_action]
)

The Lambda Action can have up to 5 inputs, and up to 5 outputs:

# fn: lambda.Function

source_output = codepipeline.Artifact()
build_output = codepipeline.Artifact()
lambda_action = codepipeline_actions.LambdaInvokeAction(
    action_name="Lambda",
    inputs=[source_output, build_output
    ],
    outputs=[
        codepipeline.Artifact("Out1"),
        codepipeline.Artifact("Out2")
    ],
    lambda_=fn
)

The Lambda Action supports custom user parameters that pipeline will pass to the Lambda function:

# fn: lambda.Function


pipeline = codepipeline.Pipeline(self, "MyPipeline")
lambda_action = codepipeline_actions.LambdaInvokeAction(
    action_name="Lambda",
    lambda_=fn,
    user_parameters={
        "foo": "bar",
        "baz": "qux"
    },
    # OR
    user_parameters_string="my-parameter-string"
)

The Lambda invoke action emits variables. Unlike many other actions, the variables are not static, but dynamic, defined by the function calling the PutJobSuccessResult API with the outputVariables property filled with the map of variables Example:

# later:
# project: codebuild.PipelineProject
lambda_invoke_action = codepipeline_actions.LambdaInvokeAction(
    action_name="Lambda",
    lambda_=lambda_.Function(self, "Func",
        runtime=lambda_.Runtime.NODEJS_12_X,
        handler="index.handler",
        code=lambda_.Code.from_inline("""
                    const AWS = require('aws-sdk');

                    exports.handler = async function(event, context) {
                        const codepipeline = new AWS.CodePipeline();
                        await codepipeline.putJobSuccessResult({
                            jobId: event['CodePipeline.job'].id,
                            outputVariables: {
                                MY_VAR: "some value",
                            },
                        }).promise();
                    }
                """)
    ),
    variables_namespace="MyNamespace"
)
source_output = codepipeline.Artifact()
codepipeline_actions.CodeBuildAction(
    action_name="CodeBuild",
    project=project,
    input=source_output,
    environment_variables={
        "MyVar": codebuild.BuildEnvironmentVariable(
            value=lambda_invoke_action.variable("MY_VAR")
        )
    }
)

See the AWS documentation on how to write a Lambda function invoked from CodePipeline.

AWS Step Functions

This module contains an Action that allows you to invoke a Step Function in a Pipeline:

import aws_cdk.aws_stepfunctions as stepfunctions

pipeline = codepipeline.Pipeline(self, "MyPipeline")
start_state = stepfunctions.Pass(self, "StartState")
simple_state_machine = stepfunctions.StateMachine(self, "SimpleStateMachine",
    definition=start_state
)
step_function_action = codepipeline_actions.StepFunctionInvokeAction(
    action_name="Invoke",
    state_machine=simple_state_machine,
    state_machine_input=codepipeline_actions.StateMachineInput.literal({"IsHelloWorldExample": True})
)
pipeline.add_stage(
    stage_name="StepFunctions",
    actions=[step_function_action]
)

The StateMachineInput can be created with one of 2 static factory methods: literal, which takes an arbitrary map as its only argument, or filePath:

import aws_cdk.aws_stepfunctions as stepfunctions


pipeline = codepipeline.Pipeline(self, "MyPipeline")
input_artifact = codepipeline.Artifact()
start_state = stepfunctions.Pass(self, "StartState")
simple_state_machine = stepfunctions.StateMachine(self, "SimpleStateMachine",
    definition=start_state
)
step_function_action = codepipeline_actions.StepFunctionInvokeAction(
    action_name="Invoke",
    state_machine=simple_state_machine,
    state_machine_input=codepipeline_actions.StateMachineInput.file_path(input_artifact.at_path("assets/input.json"))
)
pipeline.add_stage(
    stage_name="StepFunctions",
    actions=[step_function_action]
)

See the AWS documentation for information on Action structure reference.