AWS CodePipeline Construct Library
Pipeline
To construct an empty Pipeline:
# Construct an empty Pipeline
pipeline = codepipeline.Pipeline(self, "MyFirstPipeline")
To give the Pipeline a nice, human-readable name:
# Give the Pipeline a nice, human-readable name
pipeline = codepipeline.Pipeline(self, "MyFirstPipeline",
pipeline_name="MyPipeline"
)
Be aware that in the default configuration, the Pipeline
construct creates
an AWS Key Management Service (AWS KMS) Customer Master Key (CMK) for you to
encrypt the artifacts in the artifact bucket, which incurs a cost of
$1/month. This default configuration is necessary to allow cross-account
actions.
If you do not intend to perform cross-account deployments, you can disable
the creation of the Customer Master Keys by passing crossAccountKeys: false
when defining the Pipeline:
# Don't create Customer Master Keys
pipeline = codepipeline.Pipeline(self, "MyFirstPipeline",
cross_account_keys=False
)
If you want to enable key rotation for the generated KMS keys,
you can configure it by passing enableKeyRotation: true
when creating the pipeline.
Note that key rotation will incur an additional cost of $1/month.
# Enable key rotation for the generated KMS key
pipeline = codepipeline.Pipeline(self, "MyFirstPipeline",
# ...
enable_key_rotation=True
)
Stages
You can provide Stages when creating the Pipeline:
# Provide a Stage when creating a pipeline
pipeline = codepipeline.Pipeline(self, "MyFirstPipeline",
stages=[codepipeline.StageProps(
stage_name="Source",
actions=[]
)
]
)
Or append a Stage to an existing Pipeline:
# Append a Stage to an existing Pipeline
# pipeline: codepipeline.Pipeline
source_stage = pipeline.add_stage(
stage_name="Source",
actions=[]
)
You can insert the new Stage at an arbitrary point in the Pipeline:
# Insert a new Stage at an arbitrary point
# pipeline: codepipeline.Pipeline
# another_stage: codepipeline.IStage
# yet_another_stage: codepipeline.IStage
some_stage = pipeline.add_stage(
stage_name="SomeStage",
placement=codepipeline.StagePlacement(
# note: you can only specify one of the below properties
right_before=another_stage,
just_after=yet_another_stage
)
)
You can disable transition to a Stage:
# Disable transition to a stage
# pipeline: codepipeline.Pipeline
some_stage = pipeline.add_stage(
stage_name="SomeStage",
transition_to_enabled=False,
transition_disabled_reason="Manual transition only"
)
This is useful if you don’t want every executions of the pipeline to flow into this stage automatically. The transition can then be “manually” enabled later on.
Actions
Actions live in a separate package, aws-cdk-lib/aws-codepipeline-actions
.
To add an Action to a Stage, you can provide it when creating the Stage,
in the actions
property,
or you can use the IStage.addAction()
method to mutate an existing Stage:
# Use the `IStage.addAction()` method to mutate an existing Stage.
# source_stage: codepipeline.IStage
# some_action: codepipeline.Action
source_stage.add_action(some_action)
Custom Action Registration
To make your own custom CodePipeline Action requires registering the action provider. Look to the JenkinsProvider
in aws-cdk-lib/aws-codepipeline-actions
for an implementation example.
# Make a custom CodePipeline Action
codepipeline.CustomActionRegistration(self, "GenericGitSourceProviderResource",
category=codepipeline.ActionCategory.SOURCE,
artifact_bounds=codepipeline.ActionArtifactBounds(min_inputs=0, max_inputs=0, min_outputs=1, max_outputs=1),
provider="GenericGitSource",
version="1",
entity_url="https://docs.aws.amazon.com/codepipeline/latest/userguide/actions-create-custom-action.html",
execution_url="https://docs.aws.amazon.com/codepipeline/latest/userguide/actions-create-custom-action.html",
action_properties=[codepipeline.CustomActionProperty(
name="Branch",
required=True,
key=False,
secret=False,
queryable=False,
description="Git branch to pull",
type="String"
), codepipeline.CustomActionProperty(
name="GitUrl",
required=True,
key=False,
secret=False,
queryable=False,
description="SSH git clone URL",
type="String"
)
]
)
Cross-account CodePipelines
Cross-account Pipeline actions require that the Pipeline has not been created with
crossAccountKeys: false
.
Most pipeline Actions accept an AWS resource object to operate on. For example:
S3DeployAction
accepts ans3.IBucket
.CodeBuildAction
accepts acodebuild.IProject
.etc.
These resources can be either newly defined (new s3.Bucket(...)
) or imported
(s3.Bucket.fromBucketAttributes(...)
) and identify the resource that should
be changed.
These resources can be in different accounts than the pipeline itself. For example, the following action deploys to an imported S3 bucket from a different account:
# Deploy an imported S3 bucket from a different account
# stage: codepipeline.IStage
# input: codepipeline.Artifact
stage.add_action(codepipeline_actions.S3DeployAction(
bucket=s3.Bucket.from_bucket_attributes(self, "Bucket",
account="123456789012"
),
input=input,
action_name="s3-deploy-action"
))
Actions that don’t accept a resource object accept an explicit account
parameter:
# Actions that don't accept a resource objet accept an explicit `account` parameter
# stage: codepipeline.IStage
# template_path: codepipeline.ArtifactPath
stage.add_action(codepipeline_actions.CloudFormationCreateUpdateStackAction(
account="123456789012",
template_path=template_path,
admin_permissions=False,
stack_name=Stack.of(self).stack_name,
action_name="cloudformation-create-update"
))
The Pipeline
construct automatically defines an IAM Role for you in the
target account which the pipeline will assume to perform that action. This
Role will be defined in a support stack named
<PipelineStackName>-support-<account>
, that will automatically be deployed
before the stack containing the pipeline.
If you do not want to use the generated role, you can also explicitly pass a
role
when creating the action. In that case, the action will operate in the
account the role belongs to:
# Explicitly pass in a `role` when creating an action.
# stage: codepipeline.IStage
# template_path: codepipeline.ArtifactPath
stage.add_action(codepipeline_actions.CloudFormationCreateUpdateStackAction(
template_path=template_path,
admin_permissions=False,
stack_name=Stack.of(self).stack_name,
action_name="cloudformation-create-update",
# ...
role=iam.Role.from_role_arn(self, "ActionRole", "...")
))
Cross-region CodePipelines
Similar to how you set up a cross-account Action, the AWS resource object you pass to actions can also be in different Regions. For example, the following Action deploys to an imported S3 bucket from a different Region:
# Deploy to an imported S3 bucket from a different Region.
# stage: codepipeline.IStage
# input: codepipeline.Artifact
stage.add_action(codepipeline_actions.S3DeployAction(
bucket=s3.Bucket.from_bucket_attributes(self, "Bucket",
region="us-west-1"
),
input=input,
action_name="s3-deploy-action"
))
Actions that don’t take an AWS resource will accept an explicit region
parameter:
# Actions that don't take an AWS resource will accept an explicit `region` parameter.
# stage: codepipeline.IStage
# template_path: codepipeline.ArtifactPath
stage.add_action(codepipeline_actions.CloudFormationCreateUpdateStackAction(
template_path=template_path,
admin_permissions=False,
stack_name=Stack.of(self).stack_name,
action_name="cloudformation-create-update",
# ...
region="us-west-1"
))
The Pipeline
construct automatically defines a replication bucket for
you in the target region, which the pipeline will replicate artifacts to and
from. This Bucket will be defined in a support stack named
<PipelineStackName>-support-<region>
, that will automatically be deployed
before the stack containing the pipeline.
If you don’t want to use these support stacks, and already have buckets in
place to serve as replication buckets, you can supply these at Pipeline definition
time using the crossRegionReplicationBuckets
parameter. Example:
# Supply replication buckets for the Pipeline instead of using the generated support stack
pipeline = codepipeline.Pipeline(self, "MyFirstPipeline",
# ...
cross_region_replication_buckets={
# note that a physical name of the replication Bucket must be known at synthesis time
"us-west-1": s3.Bucket.from_bucket_attributes(self, "UsWest1ReplicationBucket",
bucket_name="amzn-s3-demo-bucket",
# optional KMS key
encryption_key=kms.Key.from_key_arn(self, "UsWest1ReplicationKey", "arn:aws:kms:us-west-1:123456789012:key/1234-5678-9012")
)
}
)
See the AWS docs here for more information on cross-region CodePipelines.
Creating an encrypted replication bucket
If you’re passing a replication bucket created in a different stack, like this:
# Passing a replication bucket created in a different stack.
app = App()
replication_stack = Stack(app, "ReplicationStack",
env=Environment(
region="us-west-1"
)
)
key = kms.Key(replication_stack, "ReplicationKey")
replication_bucket = s3.Bucket(replication_stack, "ReplicationBucket",
# like was said above - replication buckets need a set physical name
bucket_name=PhysicalName.GENERATE_IF_NEEDED,
encryption_key=key
)
# later...
codepipeline.Pipeline(replication_stack, "Pipeline",
cross_region_replication_buckets={
"us-west-1": replication_bucket
}
)
When trying to encrypt it (and note that if any of the cross-region actions happen to be cross-account as well, the bucket has to be encrypted - otherwise the pipeline will fail at runtime), you cannot use a key directly - KMS keys don’t have physical names, and so you can’t reference them across environments.
In this case, you need to use an alias in place of the key when creating the bucket:
# Passing an encrypted replication bucket created in a different stack.
app = App()
replication_stack = Stack(app, "ReplicationStack",
env=Environment(
region="us-west-1"
)
)
key = kms.Key(replication_stack, "ReplicationKey")
alias = kms.Alias(replication_stack, "ReplicationAlias",
# aliasName is required
alias_name=PhysicalName.GENERATE_IF_NEEDED,
target_key=key
)
replication_bucket = s3.Bucket(replication_stack, "ReplicationBucket",
bucket_name=PhysicalName.GENERATE_IF_NEEDED,
encryption_key=alias
)
Variables
Variables are key-value pairs that can be used to dynamically configure actions in your pipeline.
There are two types of variables, Action-level variables and Pipeline-level variables. Action-level variables are produced when an action is executed. Pipeline-level variables are defined when the pipeline is created and resolved at pipeline run time. You specify the Pipeline-level variables when the pipeline is created, and you can provide values at the time of the pipeline execution.
Action-level variables
The library supports action-level variables.
Each action class that emits variables has a separate variables interface,
accessed as a property of the action instance called variables
.
You instantiate the action class and assign it to a local variable;
when you want to use a variable in the configuration of a different action,
you access the appropriate property of the interface returned from variables
,
which represents a single variable.
Example:
# MyAction is some action type that produces variables, like EcrSourceAction
my_action = MyAction(
# ...
action_name="myAction"
)
OtherAction(
# ...
config=my_action.variables.my_variable,
action_name="otherAction"
)
The namespace name that will be used will be automatically generated by the pipeline construct, based on the stage and action name; you can pass a custom name when creating the action instance:
# MyAction is some action type that produces variables, like EcrSourceAction
my_action = MyAction(
# ...
variables_namespace="MyNamespace",
action_name="myAction"
)
There are also global variables available,
not tied to any action;
these are accessed through static properties of the GlobalVariables
class:
# OtherAction is some action type that produces variables, like EcrSourceAction
OtherAction(
# ...
config=codepipeline.GlobalVariables.execution_id,
action_name="otherAction"
)
The following is an actual code example.
# source_action: codepipeline_actions.S3SourceAction
# source_output: codepipeline.Artifact
# deploy_bucket: s3.Bucket
codepipeline.Pipeline(self, "Pipeline",
stages=[codepipeline.StageProps(
stage_name="Source",
actions=[source_action]
), codepipeline.StageProps(
stage_name="Deploy",
actions=[
codepipeline_actions.S3DeployAction(
action_name="DeployAction",
# can reference the variables
object_key=f"{sourceAction.variables.versionId}.txt",
input=source_output,
bucket=deploy_bucket
)
]
)
]
)
Check the documentation of the aws-cdk-lib/aws-codepipeline-actions
for details on how to use the variables for each action class.
See the CodePipeline documentation for more details on how to use the variables feature.
Pipeline-level variables
You can add one or more variables at the pipeline level. You can reference this value in the configuration of CodePipeline actions. You can add the variable names, default values, and descriptions when you create the pipeline. Variables are resolved at the time of execution.
Note that using pipeline-level variables in any kind of Source action is not supported. Also, the variables can only be used with pipeline type V2.
# source_action: codepipeline_actions.S3SourceAction
# source_output: codepipeline.Artifact
# deploy_bucket: s3.Bucket
# Pipeline-level variable
variable = codepipeline.Variable(
variable_name="bucket-var",
description="description",
default_value="sample"
)
codepipeline.Pipeline(self, "Pipeline",
pipeline_type=codepipeline.PipelineType.V2,
variables=[variable],
stages=[codepipeline.StageProps(
stage_name="Source",
actions=[source_action]
), codepipeline.StageProps(
stage_name="Deploy",
actions=[
codepipeline_actions.S3DeployAction(
action_name="DeployAction",
# can reference the variables
object_key=f"{variable.reference()}.txt",
input=source_output,
bucket=deploy_bucket
)
]
)
]
)
Or append a variable to an existing pipeline:
# pipeline: codepipeline.Pipeline
variable = codepipeline.Variable(
variable_name="bucket-var",
description="description",
default_value="sample"
)
pipeline.add_variable(variable)
Events
Using a pipeline as an event target
A pipeline can be used as a target for a CloudWatch event rule:
# A pipeline being used as a target for a CloudWatch event rule.
import aws_cdk.aws_events_targets as targets
import aws_cdk.aws_events as events
# pipeline: codepipeline.Pipeline
# kick off the pipeline every day
rule = events.Rule(self, "Daily",
schedule=events.Schedule.rate(Duration.days(1))
)
rule.add_target(targets.CodePipeline(pipeline))
When a pipeline is used as an event target, the “codepipeline:StartPipelineExecution” permission is granted to the AWS CloudWatch Events service.
Event sources
Pipelines emit CloudWatch events. To define event rules for events emitted by
the pipeline, stages or action, use the onXxx
methods on the respective
construct:
# Define event rules for events emitted by the pipeline
import aws_cdk.aws_events as events
# my_pipeline: codepipeline.Pipeline
# my_stage: codepipeline.IStage
# my_action: codepipeline.Action
# target: events.IRuleTarget
my_pipeline.on_state_change("MyPipelineStateChange", target=target)
my_stage.on_state_change("MyStageStateChange", target)
my_action.on_state_change("MyActionStateChange", target)
CodeStar Notifications
To define CodeStar Notification rules for Pipelines, use one of the notifyOnXxx()
methods.
They are very similar to onXxx()
methods for CloudWatch events:
# Define CodeStar Notification rules for Pipelines
import aws_cdk.aws_chatbot as chatbot
# pipeline: codepipeline.Pipeline
target = chatbot.SlackChannelConfiguration(self, "MySlackChannel",
slack_channel_configuration_name="YOUR_CHANNEL_NAME",
slack_workspace_id="YOUR_SLACK_WORKSPACE_ID",
slack_channel_id="YOUR_SLACK_CHANNEL_ID"
)
rule = pipeline.notify_on_execution_state_change("NotifyOnExecutionStateChange", target)
Trigger
To trigger a pipeline with Git tags or branches, specify the triggers
property.
The triggers can only be used with pipeline type V2.
Push filter
Pipelines can be started based on push events. You can specify the pushFilter
property to
filter the push events. The pushFilter
can specify Git tags.
In the case of Git tags, your pipeline starts when a Git tag is pushed.
You can filter with glob patterns. The tagsExcludes
takes priority over the tagsIncludes
.
# source_action: codepipeline_actions.CodeStarConnectionsSourceAction
# build_action: codepipeline_actions.CodeBuildAction
codepipeline.Pipeline(self, "Pipeline",
pipeline_type=codepipeline.PipelineType.V2,
stages=[codepipeline.StageProps(
stage_name="Source",
actions=[source_action]
), codepipeline.StageProps(
stage_name="Build",
actions=[build_action]
)
],
triggers=[codepipeline.TriggerProps(
provider_type=codepipeline.ProviderType.CODE_STAR_SOURCE_CONNECTION,
git_configuration=codepipeline.GitConfiguration(
source_action=source_action,
push_filter=[codepipeline.GitPushFilter(
tags_excludes=["exclude1", "exclude2"],
tags_includes=["include*"]
)]
)
)]
)
Pull request filter
Pipelines can be started based on pull request events. You can specify the pullRequestFilter
property to
filter the pull request events. The pullRequestFilter
can specify branches, file paths, and event types.
In the case of branches, your pipeline starts when a pull request event occurs on the specified branches.
You can filter with glob patterns. The branchesExcludes
takes priority over the branchesIncludes
.
# source_action: codepipeline_actions.CodeStarConnectionsSourceAction
# build_action: codepipeline_actions.CodeBuildAction
codepipeline.Pipeline(self, "Pipeline",
pipeline_type=codepipeline.PipelineType.V2,
stages=[codepipeline.StageProps(
stage_name="Source",
actions=[source_action]
), codepipeline.StageProps(
stage_name="Build",
actions=[build_action]
)
],
triggers=[codepipeline.TriggerProps(
provider_type=codepipeline.ProviderType.CODE_STAR_SOURCE_CONNECTION,
git_configuration=codepipeline.GitConfiguration(
source_action=source_action,
pull_request_filter=[codepipeline.GitPullRequestFilter(
branches_excludes=["exclude1", "exclude2"],
branches_includes=["include*"]
)]
)
)]
)
File paths can also be specified along with the branches to start the pipeline.
You can filter with glob patterns. The filePathsExcludes
takes priority over the filePathsIncludes
.
# source_action: codepipeline_actions.CodeStarConnectionsSourceAction
# build_action: codepipeline_actions.CodeBuildAction
codepipeline.Pipeline(self, "Pipeline",
pipeline_type=codepipeline.PipelineType.V2,
stages=[codepipeline.StageProps(
stage_name="Source",
actions=[source_action]
), codepipeline.StageProps(
stage_name="Build",
actions=[build_action]
)
],
triggers=[codepipeline.TriggerProps(
provider_type=codepipeline.ProviderType.CODE_STAR_SOURCE_CONNECTION,
git_configuration=codepipeline.GitConfiguration(
source_action=source_action,
pull_request_filter=[codepipeline.GitPullRequestFilter(
branches_excludes=["exclude1", "exclude2"],
branches_includes=["include1", "include2"],
file_paths_excludes=["/path/to/exclude1", "/path/to/exclude2"],
file_paths_includes=["/path/to/include1", "/path/to/include1"]
)]
)
)]
)
To filter types of pull request events for triggers, you can specify the events
property.
# source_action: codepipeline_actions.CodeStarConnectionsSourceAction
# build_action: codepipeline_actions.CodeBuildAction
codepipeline.Pipeline(self, "Pipeline",
pipeline_type=codepipeline.PipelineType.V2,
stages=[codepipeline.StageProps(
stage_name="Source",
actions=[source_action]
), codepipeline.StageProps(
stage_name="Build",
actions=[build_action]
)
],
triggers=[codepipeline.TriggerProps(
provider_type=codepipeline.ProviderType.CODE_STAR_SOURCE_CONNECTION,
git_configuration=codepipeline.GitConfiguration(
source_action=source_action,
pull_request_filter=[codepipeline.GitPullRequestFilter(
branches_excludes=["exclude1", "exclude2"],
branches_includes=["include1", "include2"],
events=[codepipeline.GitPullRequestEvent.OPEN, codepipeline.GitPullRequestEvent.CLOSED
]
)]
)
)]
)
Append a trigger to an existing pipeline
You can append a trigger to an existing pipeline:
# pipeline: codepipeline.Pipeline
# source_action: codepipeline_actions.CodeStarConnectionsSourceAction
pipeline.add_trigger(
provider_type=codepipeline.ProviderType.CODE_STAR_SOURCE_CONNECTION,
git_configuration=codepipeline.GitConfiguration(
source_action=source_action,
push_filter=[codepipeline.GitPushFilter(
tags_excludes=["exclude1", "exclude2"],
tags_includes=["include*"]
)]
)
)
Execution mode
To control the concurrency behavior when multiple executions of a pipeline are started, you can use the executionMode
property.
The execution mode can only be used with pipeline type V2.
codepipeline.Pipeline(self, "Pipeline",
pipeline_type=codepipeline.PipelineType.V2,
execution_mode=codepipeline.ExecutionMode.PARALLEL
)
Migrating a pipeline type from V1 to V2
To migrate your pipeline type from V1 to V2, you just need to update the pipelineType
property to PipelineType.V2
.
This migration does not cause replacement of your pipeline.
When the @aws-cdk/aws-codepipeline:defaultPipelineTypeToV2
feature flag is set to true
(default for new projects),
the V2 type is selected by default if you do not specify a value for pipelineType
property. Otherwise, the V1 type is selected.
codepipeline.Pipeline(self, "Pipeline",
pipeline_type=codepipeline.PipelineType.V2
)
See the CodePipeline documentation for more details on the differences between each type.