Azure AD: Managing Access to Groups

To make this example simple, you can act as both the Azure user gaining access to the system and the approver of the system. This will allow you to go through the flow of requesting and approving access from the same account. In a more common production scenario, these calls would be happening from two separate accounts.

We will be using the Azure Starter Kit as a base and replace configuration stubs for our use case.

Before you start

  1. Make sure you have:

  2. Make sure you setup:

Step 1: Configure Your Grant Kit

Configure Output

Grant Kits rely on your GitHub account and repository name to output access changes, which we'll set through Terraform local variables. Update the locals block in with your account_name and repo_name
 locals {
  account_name = "" #CHANGEME
  repo_name = "" #CHANGEME

Configure Reviewers

Workflow defines who should approve an access request.

Let's update the reviewers block by adding yourself as the reviewer by switching with the email address you use to sign into Abbey.
 resource "abbey_grant_kit" "..." {
   workflow = {
     steps = [
         reviewers = {
-          one_of = [""] # CHANGEME
+          one_of = [""]

Step 2: Configure the Azure User

In the Azure console, find your User and copy the User Principal Name. Replace the stub with it.
data "azuread_user" "dev_azure_user" {
-  user_principal_name = "" # need to use azure userPrincipalName
+  user_principal_name = ""

This will allow terraform to grab the Azure user ID to generate the grant that adds the user to the group.

Step 3: Deploy Your Starter Kit

This repo uses Azure in the Github Actions job when generating terraform resources. We'll need to install the Azure CLI. On Mac:

brew update && brew install azure-cli
az login

Then we'll need to create a set of credentials for use by Github via:

az ad sp create-for-rbac --name "myApp" --role contributor \
                            --scopes /subscriptions/{subscription-id} \

The command should output a JSON object similar to this:

    "clientId": "<GUID>", # AZURE_CLIENT_ID
    "clientSecret": "<GUID>", # AZURE_CLIENT_SECRET
    "subscriptionId": "<GUID>", # AZURE_SUBSCRIPTION_ID
    "tenantId": "<GUID>", # AZURE_TENANT_ID

Next, add the following repository secrets so Github Actions can access these credentials.





  5. ABBEY_TOKEN - API token taken from Settings > API Tokens

You can do this via Github repo page -> Settings -> Secrets and Variables -> Actions -> New Repository Secret and create with the above names. You can take the values from the JSON output of the az ad sp command.

What should it look like?

We also will need to add the correct permissions to the role we created during the az ad sp create-for-rbac command. This will allow the Azure client using the secrets in the github actions job to add the user to the desired group.

Find the role by navigating to the Azure Console , then

  1. App Registrations

  2. View all registrations in Directory

  3. Select "myApp" (the app we created as part of the az command).

  4. API Permissions

  5. Add a Permissions

  6. Microsoft Graph

  7. Add Directory.ReadWrite.All with "Grant Admin Consent for Default Directory" enabled

Your final page in permissions should look like this:

To deploy your Starter Kit, follow instructions from Deploy Your Grant Kit.

Step 5: Automate Access Management

🎉 Congratulations! Abbey is now managing permissions to your Resource for you. 🎉

You can now start requesting and approving access by following the Request Access and Approve or Deny Access Requests guides.

Abbey strives to help you automate and secure access management without being intrusive.

To that end, this Pull Request contains native Terraform HCL code using normal open source Terraform Provider libraries. It represents the permissions change. In this case, it's just a simple creation of a new Terraform Resource.

After approving the request, you should be able to see that the User has been added to the Group in the Azure console. Note that because this starter kit comes with a time-based policy by default, the user will automatically be removed after 1 hour.

Next Steps

To learn more about what Privacy Policy you can configure, try one of our Step-by-Step Tutorials.

For more information on how Abbey works, visit the Key Concepts or Grant Kits page.

Last updated