Microsoft Workloads on AWS

Convert Ansible playbooks for AMI creation into EC2 Image Builder component documents

In this post, I will show you how to automate the conversion of Ansible playbooks into EC2 Image Builder components. Automating this conversion is useful for migrations to EC2 Image Builder, where the amount of Ansible playbooks is considerable and manually creating an equivalent AWS Task Orchestrator and Executor (AWSTOE) component for each playbook becomes a time-consuming and error-prone task.

Some years ago, as a DevOps Engineer, I had written all the Windows Ansible playbooks for an Amazon Machine Image (AMI) image pipeline and the Packer template to orchestrate the images creation. This year, I migrated that solution to EC2 Image Builder and noticed that I spent most of my time on converting my playbooks into components. By automating this conversion process, I was able to avoid manual code changing, minimize mistakes, save time, and complete my migration faster. Migrating my original implementation with Packer and Ansible to EC2 Image Builder required two steps:

The second step is the focus of this blog post. As an example, I use a playbook that hardens a Windows Server 2022 AMI using the Microsoft security baseline. Although the commands I present are for Windows, you may adapt them to Linux environments as well.

Solution overview

A Python script reads an Ansible playbook YAML file, loads it as a Python dictionary, modifies it, and exports the equivalent AWSTOE component YAML file. This output file is used to create an AWSTOE component in EC2 Image Builder.


  • Basic knowledge of Ansible, EC2 Image Builder, and Python development
  • Programmatic access to an AWS account with permissions to write buckets and objects to Amazon S3 and create components on EC2 Image Builder
  • Python 3.9+
  • A code editor. In this post, I will use Visual Studio Code.

Solution walk-through

Follow these steps:

  • Step 1: Create an Ansible playbook
  • Step 2: Define the YAML component document structure
  • Step 3: Map Ansible modules and parameters to AWSTOE actions and input variables
  • Step 4: Create the Python script
  • Step 5: Add the validate and test phases, then export the YAML file
  • Step 6: Execute Python script
  • Step 7: Create an EC2 Image Builder component

Step 1: Create an Ansible playbook

First, create an Ansible playbook. This is a YAML file composed of a sequence of tasks that apply a configuration. Tasks use modules to execute pre-defined actions on a target operating system.

  1. Open Visual Studio Code and create a new file with the name playbook.yml.
  2. Paste the following code into the file.
- name: Create Microsoft Security Baseline directory
      path: C:\Installers\MicrosoftBaseline\
      state: directory
- name: Copy Microsoft Security Baseline Files
      dest: C:/Installers/MicrosoftBaseline/
- name: Unzip Custom GPO
      src: C:/Installers/MicrosoftBaseline/
      dest: C:/Installers/MicrosoftBaseline/
- name: Apply Microsoft Security Baseline
    win_shell: C:\Installers\MicrosoftBaseline\Windows-Server-2022-Security-Baseline-FINAL\Scripts\Baseline-LocalInstall.ps1 -WSNonDomainJoined

To install the Microsoft security baseline, this playbook performs the following basic tasks:

  • Creates a Folder
  • Copies a file
  • Unzips the file
  • Executes a script

Many playbooks follow the same structure, so this approach can work for other use cases, too.

Step 2: Define the YAML component document structure

An AWSTOE component is represented by a YAML component document. This YAML contains these scalars: description, name, schemaVersion, and the phases list. A component requires a minimum of three phases: build, validate, and test. Each phase has steps, which are the equivalent of Ansible tasks. As you can see in the following code, the AWSTOE component build phase contains the four Ansible tasks you created in the first step. This is where you add the converted YAML.

Step 3: Map Ansible modules and parameters to AWSTOE actions and input variables

For each Ansible module in the Ansible playbook, you map an equivalent AWSTOE action module that executes the same action. For a full list of action modules, check the list of Action modules supported by AWSTOE component manager.

To parameterize each task, Ansible uses parameters while AWSTOE uses input variables. Map each Ansible module with its parameters to a corresponding AWSTOE Action module and its input variables, as shown in Table 1.

Table 1. Mapping of Ansible modules to EC2 Image Builder components
Action Ansible(YAML) AWSTOE Component(YAML)
Create a folder win_file:
    path: "Path in host"
        path: "Path in host"
Copy file win_copy:
     src: "Path local"
    dest: "Path in host"
     source: "S3 Object URI"
destination: "Path in host"
Unzip the file win_unzip:
     path: "Path in host"
         commands: "Inline shell code"
Execute a script win_shell:
   parameter: "Inline shell code"
  commands: "Inline shell code"

Step 4: Create the Python script

  1. Open Visual Studio Code and create a file with the name
  2. Add this code to
import yaml, json

# Load the Ansible playbook as a dictionary
with open('/home/andres/devlocal/playbookToComponent/playbook.yaml') as f:
         playbook = yaml.load(f, Loader=yaml.FullLoader)

# Helper functions used to format the YAML file

def append_to_build_steps(action):
      build_steps.append({"name" : task['name'],"action":action})

def add_to_build_steps(action,var):
      if action == "ExecutePowerShell":
         inputs= []

# Loop through the Ansible Playbook and convert according to the Ansible Module - AWSTOE Action mapping table

build_steps = []
for task in playbook:
    for key in task:
        if key == "win_file":
           action = "CreateFolder"
           var = {"path":task['win_file']['path']}
        elif key == "win_copy":
           action = "S3Download"
           var = {"source":"s3://my-awstoe-component-configs/"+task['win_copy']['src'],"destination":task['win_copy']['dest']}
        elif key == "win_unzip":
           action = "ExecutePowerShell"
           var = {"commands":["Expand-Archive -LiteralPath "+task['win_unzip']['src']+" -DestinationPath c:\\Installers"]}
        elif key == "win_command":
           action = "ExecutePowerShell"
           var = {"commands":[ task['win_command']]}

Step 5: Add the validate and test phases, then export the YAML file

You can add your own validation and tests, but for the goal of this post, these phases will execute a “Hello world” statement. At the end, you will export your output to a new YAML file.

  1. Add this code at the end of
# Add the validate and test phases


    'name': 'validate',
    'steps': [{
      'name': 'HelloWorldStep',
      'action': 'ExecutePowerShell',
      'inputs': {
                 'commands': ['echo "Hello World! Validate."']

    'name': 'test',
    'steps': [{
      'name': 'HelloWorldStep',
      'action': 'ExecutePowerShell',
      'inputs': {
                 'commands': ['echo "Hello World! Test."']

awstoe_final = {'name':'WindowsServer2022SecurityBaseline',
                'description': 'Install Sec Baseline for WS2022',
                'schemaVersion': 1.0,

# Export the resulting YAML

with open('component_config.yaml', 'w') as fp:
     yaml.dump(awstoe_final, fp)

Step 6: Execute Python script

Open a command line or PowerShell prompt located in the path where you stored py (in this example, it’s C:\Temp) and execute the following commands:

C:\Temp> python -m venv .venv

C:\Temp> .venv\Scripts\activate.bat

C:\Temp> pip install pyyaml

C:\Temp> python ./

The script generates a file with the name component_config.yaml.

You need to upload this file to an existing Amazon S3 bucket. You can do it using this AWS Command Line Interface (AWS CLI) command or directly in the AWS Management Console:

C:\Temp> aws s3 cp component_config.yaml s3://your-s3-bucket --region <YOUR_AWS_REGION>

Step 7: Create an EC2 Image Builder component

First, create a file named image_component_config.json. Make sure the Amazon S3 URI points to a valid URI.

  "name": "MicrosoftSecurityBaselineWS2022",
  "semanticVersion": "1.0.0",
  "description": "MicrosoftSecurityBaselineWS2022",
  "changeDescription": "Initial Version.",
  "platform": "Windows",
  "uri": "s3://your-s3-bucket/component_config.yaml",
  "tags": {
            "App": "MicrosoftSecurityBaselineWS2022Component"

Next, create the component:

C:\Temp> aws imagebuilder create-component \
         file://image_component_config.json \
         --region <YOUR_AWS_REGION>


To avoid incurring future charges, perform the following steps:

First, list all components and retrieve your component’s ARN. To do this, run the following command, which lists all of the available components for a particular region in JSON format.

C:\Temp> aws imagebuilder list-components --region <YOUR_AWS_REGION>

Look for the component you created and copy its ARN.

    "requestId": XXXXX,
    "componentVersionList": [
            "arn": <YOUR_COMPONENT_ARN>,
            "name": "MicrosoftSecurityBaselineWS2022",
            "version": "1.0.0",
            "description": "MicrosoftSecurityBaselineWS2022",
            "platform": "Windows",
            "type": "BUILD",
            "owner": XXXXX,
            "dateCreated": XXXXX

Delete the component.

C:\Temp> aws imagebuilder delete-component --component-build-version-arn <YOUR_COMPONENT_ARN> --region <YOUR_AWS_REGION>


The conversion of Ansible playbooks into EC2 Image Builder components can be automated using Python scripts. The example I presented follows the structure of a common use case in configuration management, which can be adapted to other tools and operating systems. Customers planning a large migration to EC2 Image Builder can use this idea, scale it, and save time to put the focus on the next step, which is to create an Image Recipe that defines a base image (Windows Server 2019 Base, 2022 with SQL Server, etc.) and can combine your custom components with AWS-managed components. In the end, all resources (components, recipes, images, infrastructure configurations and distribution settings) will come together as an EC2 Image Builder Pipeline.

AWS can help you assess how your company can get the most out of cloud. Join the millions of AWS customers that trust us to migrate and modernize their most important applications in the cloud. To learn more on modernizing Windows Server or SQL Server, visit Windows on AWSContact us to start your modernization journey today.

Andres Estupinan

Andres Estupinan

Andres Estupinan is a DevOps Consultant at AWS Germany in Berlin. Before joining Amazon, he spent over 13 years working in IT across different industries such as Finance, Media and Transportation on systems infrastructure, systems administration, software development and DevOps.