Ezpresso is a CDK app that will automatically provision user-defined AWS SSO Permissions Sets and Assignments based on account structure and mappings for the AWS Organization.
Ezpresso is backed by the AWS CDK Pipeline for automated deployment of changes.
If accounts are provisioned through Control Tower or changes are made to the OU structure, the pipeline will automatically start execution through Lifecycle Events, specifically the following events:
Note that the CDK app configuration is split into two parts:
- SSO Permission Sets configuration, which can be carried between AWS Organizations. Permission Sets are configured in a YAML file, which can be checked in. This file must be checked into the repository
- Organization-specific configuration. This configuration is provided via CDK context. This configuration can be provided in one of three ways - see later.
The following diagram provides an overview of the Ezpresso architecture.
The following prerequisites must be met to carry out a successful deployment of the CDK app and pipeline.
The CDK requires the Node.js 16+ runtime.
A simple option is to use nvm to manage Node without the need to use root.
CDK version 2+ must be installed to deploy the pipeline and stacks.
The simplest option is to execute npm install -g aws-cdk through nvm-sh, without the need to use root.
Source a suitable set of AWS credentials to the AWS Organizations & Control Tower management account, which will allow deployment of AWS resources including CodePipeline, CodeBuild, IAM roles and policies.
Fork this repository to your own or a company GitHub owner/name.
Record the GitHub owner/name for later.
Create a CodeStar Connection to GitHub in the AWS Organizations Management account in the region where SSO is deployed.
This connection enables the CDK pipeline to get access to the GitHub repository and set-up a webhook for commit/merge notifications.
Record the ARN of the connection.
Bootstrap the AWS Organizations management account with the CDK in the region where AWS SSO is installed and configured:
cdk bootstrap aws://<account id>/<region>To uniquely identify groups in the chosen SSO-integrated identity store, it is strongly recommended that a prefix is chosen for each group.
This makes identifying, filtering and sorting groups in the identity store simpler.
Example: AWS_SSO_.
-
SSO instance ARN: In the AWS Organizations management account either use the AWS Identity Center Management Console to find the ARN or execute the following CLI command:
aws sso-admin list-instances --query ' Instances[0].InstanceArn' -
SSO Identity Store Id: In the AWS Organizations management account either use the AWS Identity Center Management Console to find the Id or execute the following CLI command:
aws sso-admin list-instances --query ' Instances[0].IdentityStoreId'
Record the values for later.
Before attempting to deploy SSO Permission Sets, the mapped identity store groups must exist in the chosen identity store and be replicated in the AWS SSO identity store.
Important! Ensure that all of the created identity store groups have the group prefix, e.g. AWS_SSO_, which will be used to link the SSO permission sets to the groups (see above).
SSO Permission Sets are defined in the file config/sso-permission-sets.yaml (examples provided in this file).
The SSO Permission Sets can be deployed in stages (Cloudformation stacks) on the pipeline. See Notes for use cases.
The CDK app contains a CDK pipeline and requires input parameters in the form of CDK app context.
There are several ways to provide the CDK app configuration, including:
-
--context/-coption when running the CDK app (e.g.cdk deployorcdk synth) -
cdk.jsonfile in the git repository directory -
$HOME/.cdk.jsonfile in the user's home directory
Important! Do not attempt to mix these context options.
The required configuration for the CDK app per the last two options is:
{
"context": {
"env_name": "dev", # Environment and pipeline postfix (e.g. dev/test/prod)
"permission_sets_file": "config/sso-permission-sets.yaml", # The Yaml file containing the SSO Permission Sets definitionsß
"codestar_connection_arn": "arn:aws:codestar-connections:eu-west-1:123456789012:connection/12345678-1234-1234-abcd-1234567890ab", # The CodeStar Connection Arn from above
"github_repo": "virtuability/ezpresso", # GitHub repository owner/name
"github_repo_branch": "main", # Git branch that CDK app and pipeline run from
"sso_instance_arn": "arn:aws:sso:::instance/ssoins-01234567890abcdef", # The SSO instance Arn from above
"identity_store": "d-1234567890", # The identity store id from above
"group_prefix": "AWS_SSO_" # The chosen identity store group name prefix
}
}The CDK app creates a pipeline, which is associated with the chosen GitHub repository branch (e.g. main). This enables the pipeline to automatically deploy changes on commit/merge to the branch
To deploy the CDK app & pipeline follow these instructions from a shell:
If using a cdk.json context file, run:
cdk deployIf entering input parameters directly, run:
cdk deploy \
-c env_name=dev \
-c permission_sets_file=config/sso-permission-sets.yaml \
-c codestar_connection_arn=arn:aws:codestar-connections:eu-west-1:123456789012:connection/12345678-1234-1234-abcd-1234567890ab \
-c github_repo=virtuability/ezpresso \
-c github_repo_branch=main \
-c sso_instance_arn=arn:aws:sso:::instance/ssoins-01234567890abcdef \
-c identity_store=d-1234567890 \
-c group_prefix=AWS_SSO_As previously mentioned, the CDK pipeline and CDK app will automatically execute on certain Control Tower Lifecycle Events.
In addition, the pipeline will automatically execute on commit/merge to the chosen pipeline branch.
Finally, the CDK pipeline can also be manually executed by logging in to the AWS Management Console CodePipeline page, choosing the correct region and then the pipeline and clicking on the Release Change button.
To develop, add a python virtualenv .venv and install the dependencies:
python3 -m venv .venv
source .venv/bin/activate
pip3 install -r requirements.txt
pip3 install -r requirements-dev.txtTo run CDK app tests, simply run pytest.
The following points are worth bearing in mind about the CDK app implementation:
-
The
aws identitystore list-groupsSDK & CLI call do not allow for wildcard searches. Identity store group names must therefore be predictable (lookup bygroup prefixes+group name). This is achieved in the scriptscripts/get-sso-group-mappings.py, which iterates through the SSO Permission Sets configuration file to collect all group names to perform a lookup of each SSO group id -
Several stages/stacks of SSO Permission Sets can be deployed by the pipeline. There can be several reasons for doing this:
- To logically group SSO Permission Sets and deploy them in order
- So many defined permission sets that SSO throttles Cloudformation deployment
- To ensure that the size of the Cloudformation template stays below AWS limits
cdk lslist all stacks in the appcdk synthemits the synthesized CloudFormation templatecdk deploydeploy this stack to your default AWS account/regioncdk diffcompare deployed stack with current statecdk docsopen CDK documentation
Enjoy!
Copyright 2022 Virtuability
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.