Get secure access for Azure resources in Azure Kubernetes Service by using Trusted Access

This article shows you how to get secure access for your Azure services to your Kubernetes API server in Azure Kubernetes Service (AKS) using Trusted Access.

The Trusted Access feature gives services secure access to AKS API server by using the Azure back end without requiring a private endpoint. Instead of relying on identities that have Microsoft Entra permissions, this feature can use your system-assigned managed identity to authenticate with the managed services and applications that you want to use with your AKS clusters.

Trusted Access feature overview

Trusted Access addresses the following scenarios:

• If an authorized IP range is set or in a private cluster, Azure services might not be able to access the Kubernetes API server unless you implement a private endpoint access model.
• Giving an Azure service admin access to the Kubernetes API doesn't follow the least privilege access best practice and can lead to privilege escalations or risk of credentials leakage. For example, you might have to implement high-privileged service-to-service permissions, and they aren't ideal in an audit review.

You can use Trusted Access to give explicit consent to your system-assigned managed identity of allowed resources to access your AKS clusters by using an Azure resource called a role binding. Your Azure resources access AKS clusters through the AKS regional gateway via system-assigned managed identity authentication. The appropriate Kubernetes permissions are assigned via an Azure resource called a role. Through Trusted Access, you can access AKS clusters with different configurations including but not limited to private clusters, clusters that have local accounts turned off, Microsoft Entra clusters, and authorized IP range clusters.

Prerequisites

• An Azure account with an active subscription. Create a trial subscription.

• Resource types that support system-assigned managed identity.

• Azure CLI version 2.53.0 or later. Run az --version to find your version. If you need to install or upgrade, see Install Azure CLI.

• To learn what roles to use in different scenarios, see these articles:

  • Azure Machine Learning access to AKS clusters with special configurations

• In the same subscription as the Azure resource that you want to access the cluster, create an AKS cluster.

Connect to your cluster

Configure kubectl to connect to your cluster using the az aks get-credentials command.

export RESOURCE_GROUP_NAME="myResourceGroup"
export CLUSTER_NAME="myClusterName"

az aks get-credentials --resource-group ${RESOURCE_GROUP_NAME} --name ${CLUSTER_NAME} --overwrite-existing

Verify the connection to your cluster using the kubectl get command.

kubectl get nodes

Select the required Trusted Access roles

The roles that you select depend on the Azure services that you want to access the AKS cluster. Azure services help create roles and role bindings that build the connection from the Azure service to AKS.

To find the roles that you need, see the documentation for the Azure service that you want to connect to AKS. You can also use the Azure CLI to list the roles that are available for the Azure service using the az aks trustedaccess role list --location <location> command.

Create a Trusted Access role binding

After you confirm which role to use, use the Azure CLI to create a Trusted Access role binding in the AKS cluster. The role binding associates your selected role with the Azure service.

export ROLE_BINDING_NAME="myRoleBindingName"
export SOURCE_RESOURCE_ID="mySourceResourceID"
export ROLE_NAME_1="myRoleName1"
export ROLE_NAME_2="myRoleName2"

az aks trustedaccess rolebinding create --resource-group ${RESOURCE_GROUP_NAME} --cluster-name ${CLUSTER_NAME} --name ${ROLE_BINDING_NAME} --source-resource-id ${SOURCE_RESOURCE_ID} --roles ${ROLE_NAME_1},${ROLE_NAME_2}

Results:

{
  "id": "/subscriptions/aaaa0a0a-bb1b-cc2c-dd3d-eeeeee4e4e4e/resourceGroups/${RESOURCE_GROUP_NAME}/providers/Microsoft.ContainerService/managedClusters/${CLUSTER_NAME}/trustedAccessRoleBindings/${ROLE_BINDING_NAME}",
  "name": "${ROLE_BINDING_NAME}",
  "provisioningState": "Succeeded",
  "resourceGroup": "${RESOURCE_GROUP_NAME}",
  "roles": [
    "${ROLE_NAME_1}",
    "${ROLE_NAME_2}"
  ],
  "sourceResourceId": "${SOURCE_RESOURCE_ID}",
  "systemData": null,
  "type": "Microsoft.ContainerService/managedClusters/trustedAccessRoleBindings"
}

Update an existing Trusted Access role binding

For an existing role binding that has an associated source service, you can update the role binding with new roles using the az aks trustedaccess rolebinding update --resource-group $RESOURCE_GROUP_NAME --cluster-name $CLUSTER_NAME --name $ROLE_BINDING_NAME --roles $ROLE_NAME_3,$ROLE_NAME_4 command. This command updates the role binding with the new roles that you specify.

Show a Trusted Access role binding

Show a specific Trusted Access role binding using the az aks trustedaccess rolebinding show --name $ROLE_BINDING_NAME --resource-group $RESOURCE_GROUP_NAME --cluster-name $CLUSTER_NAME command.

List all the Trusted Access role bindings for a cluster

List all the Trusted Access role bindings for a cluster using the az aks trustedaccess rolebinding list --resource-group $RESOURCE_GROUP_NAME --cluster-name $CLUSTER_NAME command.

Related content

• Deploy and manage cluster extensions for AKS
• Deploy the Azure Machine Learning extension on an AKS–enabled Kubernetes cluster