AWS Cloud Operations & Migrations Blog

Using AWS CodePipeline to deploy AWS Config conformance packs created with the Rule Development Kit

As consultants, we often help customers manage AWS services using infrastructure as code (IaC). We follow DevOps practices for building, versioning, testing, and deploying services. We also use AWS Config custom and managed rules to evaluate the configuration settings of AWS resources. AWS Config continuously tracks the configuration changes that occur among AWS resources and checks whether these changes violate any of the conditions in the rules. If a resource violates a rule, AWS Config flags the resource and the rule as non-compliant. This approach helps us demonstrate compliance with our customers’ internal company policies.

We wanted to deploy AWS Config rules across multiple accounts in an organization created in AWS Organizations in a scalable, consistent, and efficient way, aligned with our DevOps practices.

In this blog post, we share a sample implementation to build, test, and deploy AWS Config rules at scale across multiple AWS accounts. We use AWS Config conformance packs, the Rule Development Kit (RDK), RDKlib, and the services listed in the following table. The implementation packages AWS Config rules created with the RDK and RDKlib into an organization conformance pack template. CodeBuild deploys the template. This implementation also supports sample conformance pack templates provided by AWS.


Figure 1 shows the architecture of the implementation.


Check the code into CodeCommit. CodePipeline pipeline starts and triggers a CodeBuild job. Build artifacts are sent to an S3 bucket. The conformance packs composed of AWS Config rules and Lambda functions are deployed using CloudFormation.

Figure 1: Architecture of AWS Config conformance packs pipeline

The implementation is an automated pipeline that uses CodeCommit to host the code for AWS Config custom rules. The code was created using the RDK and conformance pack template files in YAML format. CodeBuild deploys AWS Lambda functions for the AWS Config custom rules. It packages the rules into a conformance pack template. CodeBuild deploys all conformance pack templates located at the root level of the CodeCommit repository, including the ones created using the RDK. CodePipeline automatically releases changes made to the content stored in the CodeCommit repository.


Before you begin:

1. Install the AWS CLI. For instructions, see Installing the AWS CLI in the AWS Command Line Interface User Guide.

2. Complete the prerequisites in the Deploy Conformance Packs Across an Organization with Automatic Remediation blog post.

3. Install the RDK and RDKlib locally:

# Using pip
pip install rdk
pip install rdklib

4. In the management account, install the RDKlib layer from the AWS Serverless Repository as described in the “Install the RDKlib layer” section of  the AWS Config Rule Development Kit library: Build and operate rules at scale blog post. Make a note of the ARN of the deployed layer. You will use it later in this post.

5. (Optional) If you want to deploy and manage conformance packs from an AWS Organizations non-management account, follow the instructions in the Deploy AWS Config rules and conformance packs using a delegated admin blog post.

6. In the management account, make a note of the root ID for the organization. It starts with r- and can be found using the console or AWS CLI.

  • In the AWS Organizations console, choose the Organize accounts tab. In the left pane, choose Root, and from the right pane, look under ARN.
  • Use the following AWS CLI command:
aws organizations list-roots --query "Roots[0].Id"
# arn:aws:organizations::123456789012:root/o-8snfo67u2v/r-i2ey


To deploy the implementation:

# Clone the repository
git clone

# Move to the repository's directory
cd aws-config-conformance-packs-pipeline

Create an IAM role for the Lambda function to verify rule compliance on every account. For simplicity, we included a CloudFormation template that can be deployed as a stackset on all the accounts that you want to monitor. To deploy the stackset using the method described in this post, enable trusted access with AWS Organizations for AWS CloudFormation StackSets. You can also deploy the stackset using self-managed permissions, but you will need to adapt the following commands accordingly.

Our example uses the eu-west-1 AWS Region. First, set up your AWS credentials. Replace <ManagementAccountId> with the ID of your management account and <OrganizationalUnitId> with the value from step 6 in the “Prerequisites” section.

# Deploy the IAM role used by the RDK Lambda function
aws cloudformation create-stack-set --stack-set-name RDKLambdaRole --template-body file://rdk-role.yml --parameters ParameterKey=LambdaAccountId,ParameterValue=<ManagementAccountId> --capabilities CAPABILITY_NAMED_IAM --permission-model SERVICE_MANAGED --auto-deployment Enabled=true,RetainStacksOnAccountRemoval=false
aws cloudformation create-stack-instances --stack-set-name RDKLambdaRole --regions eu-west-1 --deployment-targets OrganizationalUnitIds=<OrganizationalUnitId>

Next, deploy the pipeline. For <RDKlibLayerArn>, use the value from step 4 of the “Prerequisites” section.

# Deploy the AWS CloudFormation template, don’t forget to replace <RDKlibLayerArn> with the value you noted on step 4 of the prerequisites
aws cloudformation deploy --template-file cfn-template.yml --stack-name org-conformance-packs-pipeline --parameter-overrides RDKlibLayerArn=<RDKlibLayerArn> --capabilities CAPABILITY_NAMED_IAM

After the deployment is complete, you can use the solution.


To use the RDK to create a custom AWS Config rule, see the RDK Workshop or the AWS Config Rule Development Kit library: Build and operate rules at scale blog post.

First, use your CodeCommit credentials to set up your local Git repository. For instructions, see Setting up for AWS CodeCommit in the AWS CodeCommit User Guide.

To use the solution to deploy AWS Config custom rules created using the RDK or conformance pack templates:

1. In your local machine, move to the repository’s directory.

2. Create an AWS Config custom rule or copy an existing one from the AWS Config rules repository.

a. For a periodic trigger:

# Replace <maximum frequency> with your desired frequency
rdk create <rulename> –-runtime python3.6-lib --maximum-frequency <maximum frequency>

For information about the maximum frequency supported by AWS Config, see MaximumExecutionFrequency in the AWS CloudFormation User Guide.

b. For a configuration change trigger:

# Replace <resource type> with your desired resource type
rdk create YOUR_RULE_NAME --runtime python3.6-lib --resource-types <resource_type>

For information about resource types supported by AWS Config, see Supported Resource Types in the AWS Config Developer Guide.

3. Add the name of the IAM role you created using the stackset to the parameters.json file in the rule directory. The Lambda function assumes this role to verify that the resources on the target account are compliant.

"Version": "1.0",
"Parameters": {
"RuleName": "test_rule",
"Description": "test_rule",
"SourceRuntime": "python3.6-lib",
"CodeKey": "",
"InputParameters": "{\"ExecutionRoleName\": \"AWSRDKLambdaRole\"}",
"OptionalParameters": "{}",
"SourceEvents": "AWS::EC2::Instance"
"Tags": "[]"

4. Create an AWS Config managed conformance pack template or copy an existing one from sample conformance packs templates provided by AWS.

5. Push your code to the remote CodeCommit repository.

There are two stages in the CodePipeline pipeline: a CodeCommit source stage and CodeBuild build stage. The source stage is triggered whenever code is pushed to the CodeCommit repository. The build stage copies the code to an S3 staging bucket using CodeBuild. The build stage is defined in a buildspec.yaml file that uses the AWS CLI to:

  • Set up the RDK in the build environment.
  • Deploy Lambda functions for AWS Config custom rules created using the RDK.
  • Create a conformance pack template file named custom-conforms-pack-template.yaml.
  • Deploy conformance pack templates using the aws configservice put-organization-conformance-pack command.

You have now deployed the AWS Config conformance packs. You should be able to see them on the Conformance Packs page in the AWS Config console.


To avoid incurring additional charges in your account, run the following commands in the management account or the delegated admin account to remove the resources you created:

# Delete all conformance packs deployed

aws configservice delete-organization-conformance-pack
--organization-conformance-pack-name &lt;value&gt;

# Delete all objects created in the S3 bucket used to store CodePipeline artifacts
aws s3 rm s3://&lt;bucket_name&gt;/ --recursive

# Delete the AWS CloudFormation stack
aws cloudformation delete --stack-name org-conformance-packs-pipeline

Delete the resources you created as prerequisites for deploying conformance packs across an organization.


In this post, we shared an implementation capable of defining compliance rules in a centralized manner using the RDK and deploying these rules across multiple accounts using AWS CodePipeline and AWS CodeBuild.

The implementation offers the following advantages:

  • It follows a DevOps workflow and decouples compliance rule development, review, and deployment.
  • Through the use of CodeCommit, it facilitates change management of compliance rules.
  • It scales well with an increasing number of AWS accounts and centralizes compliance management when using a distributed organization model.

To download the CloudFormation template, see the dedicated code repository on GitHub. For information about pricing, see AWS Config pricing, AWS CodeBuild pricing, and AWS CodePipeline pricing. For more learning resources, visit the AWS Management and Governance Blog.

About the authors

Sebastian Caceres

Sebastian Caceres is a DevOps Team Leader in Professional Services at Amazon Web Services. He helps customers solve problems related to automation, cloud architecture and infrastructure, development, operations, and organization.

Mostefa Brougui

Mostefa Brougui is a Security Team Leader in Professional Services at Amazon Web Services. He works with AWS enterprise customers to design, build, and optimize their security architecture to drive business outcomes.