Simple AWS Image Pipeline

This simplified CloudFormation template creates an automated AMI building pipeline using AWS CodePipeline and CodeBuild, with support for Packer, Ansible, and Goss testing.

Overview

This implementation consolidates all resources into a single CloudFormation template, making it easier to understand and maintain compared to nested stacks. The pipeline pulls source artifacts from S3 buckets and orchestrates the build process using CodeBuild.

Included Resources

IAM roles and policies for CodeBuild and CodePipeline
S3 bucket for pipeline artifacts
CodeBuild project for building AMIs
CodePipeline with Source and Build stages

Prerequisites

An AWS account with permission to create the resources in this template
An existing S3 bucket containing your source artifacts:
- Packer templates
- Ansible playbooks
- Goss test specifications
- Pip configuration (optional)
A VPC with a private subnet and security group for CodeBuild to use

Usage

1. Prepare Your Source Artifacts

Organize your Packer templates, Ansible playbooks, and Goss tests, and upload them to your S3 bucket:

image-pipeline-packer.zip: Contains Packer templates and configuration
image-pipeline-ansible-playbooks.zip: Contains Ansible playbooks and roles
image-pipeline-goss.zip: Contains Goss test files
image-pipeline-pip-config.zip: Contains pip configuration (optional)

2. Deploy the CloudFormation Stack

Using the AWS Console:

Go to CloudFormation in the AWS Console
Click "Create stack" > "With new resources"
Upload the template.yaml file
Fill in the parameter values
Review and create the stack

Using the AWS CLI:

aws cloudformation create-stack \
  --stack-name my-image-pipeline \
  --template-body file://template.yaml \
  --parameters \
      ParameterKey=ProjectName,ParameterValue=my-image-pipeline \
      ParameterKey=VpcId,ParameterValue=vpc-12345678 \
      ParameterKey=SubnetId,ParameterValue=subnet-12345678 \
      ParameterKey=SecurityGroupId,ParameterValue=sg-12345678 \
      ParameterKey=AssetsBucketName,ParameterValue=my-assets-bucket \
  --capabilities CAPABILITY_NAMED_IAM

3. Parameters

Parameter	Description	Default
ProjectName	Unique name for the project	(Required)
BuilderImage	Docker image to use for CodeBuild	aws/codebuild/standard:7.0
BuilderComputeType	CodeBuild instance size	BUILD_GENERAL1_SMALL
VpcId	VPC ID where CodeBuild will run	(Required)
SubnetId	Subnet ID where CodeBuild will run	(Required)
SecurityGroupId	Security group ID for CodeBuild	(Required)
AssetsBucketName	S3 bucket containing source assets	image-pipeline-assets-dev
PackerS3Key	S3 key for Packer templates	image-pipeline-packer.zip
AnsibleS3Key	S3 key for Ansible playbooks	image-pipeline-ansible-playbooks.zip
GossS3Key	S3 key for Goss tests	image-pipeline-goss.zip
PipS3Key	S3 key for pip configuration	image-pipeline-pip-config.zip
SSHUser	SSH username for build instances	ec2-user

Differences from the Original Implementation

This simplified implementation:

Uses a single CloudFormation template instead of nested stacks
Removes KMS encryption for simplicity (add if needed)
Streamlines IAM permissions to only what's necessary
Simplifies the CodeBuild project configuration
Focuses exclusively on AMI building functionality
Has a simpler deployment process

Customizing Your Implementation

Adding KMS Encryption

If you need KMS encryption for your artifacts, you can add a KMS key resource and update the S3 bucket and CodePipeline configurations to use it.

Modifying the Build Process

The build process is defined in the BuildSpec of the CodeBuild project. You can customize it to fit your specific needs by modifying the commands in each phase.

Adding a Test Stage

If you want to add a dedicated test stage, you can create another CodeBuild project and add it as another stage in the CodePipeline.

CSVD/aws-image-pipeline-cfn-simple

About

Resources

Stars

Watchers

Forks

Releases

Languages