AWS CDK Python - Overriding the resource logical IDs

readme.md

Install the tools

pyenv install 3.8.1
pyenv local 3.8.1
brew install node
npm install -g aws-cdk@1.31.0

Setup the project

mkdir my_python_sample
cd my_python_sample
cdk init --language python sample-app
source .env/bin/activate
pip install -r requirements.txt

Replace local app.py and my_python_sample_stack.py files with the gists

Synthesise the CloudFormation template

cdk synth

Update the code to use the following and examine the synthesised output:

• UseOriginalConstructID aspect
• reset_logical_id()
• or override_logical_id()

app.py

#!/usr/bin/env python3

from aws_cdk import core

from my_python_sample.my_python_sample_stack import MyPythonSampleStack


app = core.App()
MyPythonSampleStack(app, 'my-python-sample', env={'region': 'ap-southeast-2'})

app.synth()

my_python_sample_stack.py

from aws_cdk import (
    aws_s3 as s3,
    core
)
import jsii

def override_logical_id(resource, logical_id):
    """ Overrides the resource logical ID with a specific ID
    """
    resource.node.default_child.override_logical_id(logical_id);

def reset_logical_id(resource):
    """ Resets the resource logical ID to the original construct ID
    """
    resource.node.default_child.override_logical_id(resource.node.id);

@jsii.implements(core.IAspect)
class UseOriginalConstructID:

  def visit(self, element):
      """ Traverses all the Construct's elements and resets their resource logical ID to the original construct ID

      Note:
        * This method works best for all explicitly defined resources
        * Implicit dependent resources created by the CDK may contain hash based or conflicting IDs
      """
      if isinstance(element.node.default_child, core.Construct):
          element.node.default_child.override_logical_id(element.node.id)


class MyPythonSampleStack(core.Stack):

    def __init__(self, scope: core.Construct, id: str, **kwargs) -> None:
        super().__init__(scope, id, **kwargs)

        bucket = s3.Bucket(self, 'MyBucket', bucket_name='my-bucket-name')
        # reset_logical_id(bucket)
        # override_logical_id(bucket, 'S3BucketResource')
        
        # Apply the UseOriginalConstructID apsect to correct all the resource logical ids
        # self.node.apply_aspect(UseOriginalConstructID())

outputs.yml

# Auto generated logical ids
Resources:
  MyBucketF68F3FF0:
    Type: AWS::S3::Bucket
    Properties:
      BucketName: my-bucket-name
    UpdateReplacePolicy: Retain
    DeletionPolicy: Retain
    Metadata:
      aws:cdk:path: my-python-sample/MyBucket/Resource
  CDKMetadata:
    Type: AWS::CDK::Metadata
    Properties:
      Modules: aws-cdk=1.31.0,...

# Resetting the logical id to the original construct id
Resources:
  MyBucket:
    Type: AWS::S3::Bucket
    Properties:
      BucketName: my-bucket-name
    UpdateReplacePolicy: Retain
    DeletionPolicy: Retain
    Metadata:
      aws:cdk:path: my-python-sample/MyBucket/Resource
  CDKMetadata:
    Type: AWS::CDK::Metadata
    Properties:
      Modules: aws-cdk=1.31.0,...

# Overriding the logical id with a custom value
Resources:
  S3BucketResource:
    Type: AWS::S3::Bucket
    Properties:
      BucketName: my-bucket-name
    UpdateReplacePolicy: Retain
    DeletionPolicy: Retain
    Metadata:
      aws:cdk:path: my-python-sample/MyBucket/Resource
  CDKMetadata:
    Type: AWS::CDK::Metadata
    Properties:
      Modules: aws-cdk=1.31.0,...

that's a super-clean example, thank you for summarizing this

Great example!!! I am wondering how to override the stack name as resources still get named using stackName-OverridenReourceName-Randomhash.