APT Lab Terraform

APT

Applied Purple Teaming Threat Optics Lab – Azure TerraForm

Purple Teaming Attack & Hunt Lab – TerraForm


Background

Defensive Origins uses a highly verbose threat optics lab to isolate adversarial techniques to more easily attribute IOC (indicators of compromise). These labs have routinely been time consuming to build and manage. The platform included here automates much of the threat-optic lab environment built on the Azure cloud network.

Installation:

This process requires Python3.

Install AzureCLI

https://docs.microsoft.com/en-us/cli/azure/install-azure-cli-apt?view=azure-cli-latest

Install TerraForm

https://learn.hashicorp.com/terraform/getting-started/install.html

Setup and Configuration

Create Token/Document

Default Credentials:

Default credentials are set in LabBuilder.py.

  • Windows & Linux systems:
itadmin:APTClass!
  • Kibana:
helk:hunting

Changing Default Credentials

The credentials can be changed within the locals variable:

locals {
  resource_group_name   = "class-resources"
  master_admin_username ="itadmin"
  master_admin_password ="APTClass!"
  master_domain         ="labs.local"
}

The password for Kibana can be changed by editing the HELK install line:

./helk_install.sh -p hunting -i 10.10.98.20 -b 'helk-kibana-analysis-alert'

within the

3-setup.tf`

This file is located at:

APT-Lab-Terraform/master/modules/linux/3-setup.tf

Please note the following regarding access:

  • Only the Windows client is accessible externally
  • Kibana is accessed internally, use a browser on one of the Windows machines to access.
  • An SSH client will need to be installed on the Windows machines in order to SSH to the Linux system.

Configure Regions

Clone APT TerraForm Repository

git clone https://github.com/DefensiveOrigins/APT-Lab-Terraform.git
cd APT-Lab-Terraform

Execute TerraForm build process

Run the builder and deploy your systems.

python .\LabBuilder.py -m YOURPUBLICIP

Please note:

  • If this script errors, or there are missing dependencies ensure it is being executed with Python 3.X. As such, attempt to execute with Python3 directly:
python3 .\LabBuilder.py -m YOURPUBLICIP

Source IP Filtering

The -m flag will accept a single IP Address or Subnet as input. This adds the IP as a SRC IP address filter on the lab environment.

-m [IP]

Successful Deployment

To confirm successful deployment the following 3 virtual machines will be found within Azure:

  • stu-client
  • stu-dc
  • stu-linux
azure

Deployment, include all the post-installation scripts, may take twenty minutes or more. Setup of the Linux node, with its ELK stack, will take the longest.

Troubleshooting Steps

If LabBuilder.py errors during execution. Delete the LABS folder, found at

APT-Lab-Terraform/LABS/

The error (shown at the top of script execution) is:

Directory not copied. Error: [Errno 17] File exists: './LABS'.

Errors referencing a ‘duplicate’ are also solved by this. Examples include:

Error: Duplicate module call
Error: Duplicate resource "azurerm_resource_group" configuration

If LabBuilder.py gives you an error as follows:

TypeError: replace() argument 2 must be str, not tuple

… it means that you copy-pasted configuration data for your Azure Cloud authentication, without removing the trailing comma (,). Definitions such as “client_id” should not end with a comma.

If LabBuilder.py creates only one or two out of the three VMs it will likely also throw an error including the following text:

Operation could not be completed as it results in exceeding approved Total Regional Cores quota.

This occurs when you have a free/trial Azure Cloud subscription, which is limited to 4 active CPU cores. You may edit the VM definitions for the Active Directory server and the Windows client, to change the VM sizing. This is done in the files named “2-virtual-machine.tf”, by replacing the “vm_size” field. The files include an example line to use as replacement.

Destroying the Lab

python .\LabBuilder.py -destroy

The ‘-d’ or ‘-destroy’ flag will execute theTerraform destroy command. This will remove the Lab in Azure. CAUTION: All data within the VMs will be deleted.

Please confirm within the Azure portal that everything has been deleted.

Accessing HELK:


Modules

The various components of this build process are defined below.

ModuleFunction
/master/modulesVarious TerraForm modules
LabBuilder.pyPython script that uses TerraForm and AzureCLI to build the Applied Purple Teaming to specifications using the modules in /master/modules and additional resources.
labs.zipAdditional resources to configure lab environment.

Network

This module creates a Network with 2 x Subnets:

  • Domain Controllers
  • Domain Clients

This module shouldn’t be used as-is in a Production Environment – where you’d probably have Network Security Rules configured – it’s designed to be a simplified configuration for the purposes of this example.

ModuleFunction
main.tfSetup Primary Network
outputs.tfGrab and Set Network ID
variables.tfTerraForm variables

Active Directory

ModuleFunction
1-network-interface.tfSetup interface of Domain Controller
2-virtual-machine.tfSpecify Domain Controller VM Attributes
3-provision-domain.tfInitial Domain Configuration
4-wait-for-domain-to-provision.tfQuietly wait for Domain to finish provisioning
5-setup.tfDomain Controller Services and Software Configuration
variables.tfTerraForm variables
files/FirstLogonCommands.xmlRun first commands
files/winrm.ps1Enable WinRM, grab Lab resources

Linux / Helk

  • This module configures a Ubuntu image with the necessary tooling to be used as a hunters-SIEM (HELK)
ModuleFunction
1-network-interface.tfSetup interface for Linux system
2-virtual-machine.tfSpecify VM configuration for Linux System
3-setup.tfSetup and configure software
variables.tfTerraForm variables

Windows Client

  • This module provisions a Windows Client which will be bound to the Active Directory Domain created in the other module.
  • There’s a few hacks in here as we have to wait for Active Directory to become available, but this takes advantage of the azurerm_virtual_machine_extension resource. It’s worth noting that the keys in this resource are case sensitive.
ModuleFunction
1-network-interface.tfSetup network interface for Windows client
2-virtual-machine.tfSpecify client VM configuration
3-wait-for-domain-to-provision.tfPolitely wait for the Domain to be provisioned
4-join-domain.tfJoin Windows client to domain.
5-setup.tfProcure and configure the various tools.
outputs.tfGrab the associated VM IP.
variables.tfTerraForm variables

Hat-Tips and Acknowledgments