RDKlib is a Python library to enable you to run custom AWS Config Rules at scale. The library can be used to:
- Help you to focus only on the compliance logic, while the library does the heavy lifting
- Ease maintenance by moving the boilerplate code as a AWS Lambda Layer
- Ease deployment by using AWS Serverless Application Repository
RDKLib works in synergy with the AWS Config Rule Development Kit.
pip install rdklib
Note: you need to install the RDK first.
To use rdklib
, specify a python3.x-lib
runtime when you run rdk create
(or don't specify any runtime; rdklib
is now the default for rdk create
). This will populate the rdklib
runtime in the RDK parameters.json
of your Rule template. Examples:
- For periodic trigger:
rdk create YOUR_RULE_NAME --runtime python3.12-lib --maximum-frequency TwentyFour_Hours
- For configuration change trigger (for example S3 Bucket):
rdk create YOUR_RULE_NAME --runtime python3.12-lib --resource-types AWS::S3::Bucket
After you've created your rule, update the .py
file that was generated, adding your custom logic within the evaluate_change()
method for change-triggered rules or the evaluate_periodic()
method for periodic rules (you may need to uncomment evaluate_periodic()
. If you need to create a boto3
client, use the client_factory
helper (eg. instead of boto3.client("s3")
, use client_factory.build_client("s3")
). Examples of rdklib
rules can be found here.
RDKlib is designed to work as a AWS Lambda Layer. It allows you to use the library without needing to include it in your deployment package.
- Install RDKlib layer (with AWS CLI)
aws serverlessrepo create-cloud-formation-change-set --application-id arn:aws:serverlessrepo:ap-southeast-1:711761543063:applications/rdklib --stack-name RDKlib-Layer
# Copy/paste the full change-set ARN to customize the following command
aws cloudformation execute-change-set --change-set-name NAME_OF_THE_CHANGE_SET
aws cloudformation describe-stack-resources --stack-name serverlessrepo-RDKlib-Layer
# Copy the ARN of the Lambda layer in the "PhysicalResourceId" key (i.e. arn:aws:lambda:YOUR_REGION:YOUR_ACCOUNT:layer:rdklib-layer:1).
Note: You can do the same step manually going to https://console.aws.amazon.com/lambda/home#/create/function?tab=serverlessApps and find "rdklib"
- Deploy the rule
rdk deploy YOUR_RULE_NAME --rdklib-layer-arn YOUR_RDKLIB_LAYER_ARN
- Q. What is the
client_factory
that I see in myrdklib
rules?- A. A
client_factory
is a class that allows for dynamic provisioning of aboto3
client. In anrdklib
rule, you should treatclient_factory
as the way to create aboto3
client. So instead of callingclient = boto3.client("s3")
, you would callclient = client_factory.build_client("s3")
.- Q. ...Why?
- A. It's mainly there to allow for cross-account functionality so that your client evaluates the rule in the right account.
- Q. ...Why?
- A. A
This project is licensed under the Apache-2.0 License.
Feel free to email [email protected]
- Benjamin Morris - Maintainer, code, testing
- Mark Beacom - Maintainer, code, testing
- Michael Borchert - Design, code, testing, feedback
- Ricky Chau - Maintainer, code, testing
- Julio Delgado Jr. - Design, testing, feedback
- Chris Gutierrez - Design, feedback
- Joe Lee - Design, feedback
- Jonathan Rault - Maintainer, design, code, testing, feedback
- Carlo DePaolis - Maintainer, code, testing