
aws-deployer
Serverless CloudFormation deployment automation with support for single and multi-account deployments via AWS Step Functions
- version
- 1.0.0
Install
/plugin install aws-deployer@savaki-marketplaceRun it inside Claude Code after adding the marketplace with /plugin marketplace add savaki/savaki-marketplace.
Contents
Discovered from sourceSkills 2
Set Up GitHub Actions Deployment for AWS Deployer
This skill automatically configures a GitHub repository to deploy via AWS Deployer (github.com/savaki/aws-deployer).
Prerequisites
- AWS Deployer infrastructure is already set up (use
setup-deployerskill if not) - Working directory is a Git repository with a GitHub remote
- Repository has a
cloudformation.templatefile (will be created if missing) - Optional:
S3_ARTIFACT_BUCKETenvironment variable set to your artifacts bucket - Optional:
AWS_REGIONenvironment variable set to your AWS region
Your Task
When this skill is invoked, you will configure the repository for GitHub Actions deployment by:
- Detecting repository information - Extract repo name from
.git/config - Getting configuration - Check environment variables (
S3_ARTIFACT_BUCKET,AWS_REGION) first, ask only if not set - Detecting build system - Identify build command from project files
- Creating/updating GitHub workflow - Generate
.github/workflows/deploy.yml - Updating CloudFormation template - Ensure required parameters exist
- Creating parameter files - Generate base and environment-specific templates
- Explaining OIDC setup - Provide instructions for running
aws-deployer setup-githuband configuring GitHub secrets
Important:
- Check environment variables first:
S3_ARTIFACT_BUCKETandAWS_REGION - Only ask for values if environment variables are not set
- Default AWS region to
us-west-2if not set - Do NOT run the
aws-deployer setup-githubcommand. Explain to the user how to run it. - Explain that THREE GitHub secrets are required:
AWS_ROLE_ARN,S3_ARTIFACT_BUCKET, andAWS_REGION.
Step 1: Detect Repository Information
Extract the repository owner and name from Git configuration:
# Get the remote URL
git config --get remote.origin.url
Parse the result to extract owner/repo:
https://github.com/foo/bar.git→foo/bargit@github.com:foo/bar.git→foo/bar
Store as REPO_FULL_NAME (e.g., foo/bar)
Step 2: Get Configuration
S3 Bucket:
- Check if
S3_ARTIFACT_BUCKETenvironment variable is setecho $S3_ARTIFACT_BUCKET - If set, use that value
- If not set, ask the user for the S3 artifacts bucket name
- Example:
my-artifacts-bucket - This will be stored as a GitHub secret
S3_ARTIFACT_BUCKET
- Example:
AWS Region:
- Check if
AWS_REGIONenvironment variable is setecho $AWS_REGION - If set, use that value
- If not set, ask the user to confirm the AWS region, providing options:
- Default:
us-west-2 - Other common options:
us-east-1,us-east-2,eu-west-1,ap-southeast-1 - Allow custom input
- Default:
Initial Environment: Default to dev (don't ask)
Step 3: Detect Build Command
Check for common build files and infer the build command:
Makefilewithbuildtarget →make buildpackage.jsonwithbuildscript →npm run buildgo.mod→go build ./...pom.xml→mvn packagebuild.gradle→gradle build
If multiple or none found, ask the user for the build command.
Step 4: Create/Update GitHub Workflow
Create .github/workflows/deploy.yml:
name: Deploy to AWS
on:
push:
branches:
- main
- develop
workflow_dispatch:
permissions:
id-token: write
contents: read
jobs:
deploy:
runs-on: ubuntu-latest
steps:
- name: Checkout code
uses: actions/checkout@v4
- name: Configure AWS credentials
uses: aws-actions/configure-aws-credentials@v4
with:
role-to-assume: ${{ secrets.AWS_ROLE_ARN }}
aws-region: ${{ secrets.AWS_REGION }}
- name: Set version variables
id: version
run: |
REPO="<repo-owner>/<repo-name>"
BRANCH="${{ github.ref_name }}"
SHA_SHORT="${{ github.sha }}"
SHA_SHORT="${SHA_SHORT:0:6}"
VERSION="${{ github.run_number }}.${SHA_SHORT}"
S3_PREFIX="${REPO}/${BRANCH}/${VERSION}"
echo "repo=${REPO}" >> $GITHUB_OUTPUT
echo "branch=${BRANCH}" >> $GITHUB_OUTPUT
echo "version=${VERSION}" >> $GITHUB_OUTPUT
echo "s3_prefix=${S3_PREFIX}" >> $GITHUB_OUTPUT
- name: Build application
run: |
<build-command>
- name: Generate CloudFormation parameters
run: |
cat > cloudformation-params.json <<EOF
{
"Env": "dev",
"Version": "${{ steps.version.outputs.version }}",
"S3Bucket": "${{ secrets.S3_ARTIFACT_BUCKET }}",
"S3Prefix": "${{ steps.version.outputs.s3_prefix }}"
}
EOF
- name: Upload to S3
run: |
S3_PATH="s3://${{ secrets.S3_ARTIFACT_BUCKET }}/${{ steps.version.outputs.s3_prefix }}/"
# Upload CloudFormation template
aws s3 cp cloudformation.template "${S3_PATH}"
# Upload parameters
aws s3 cp cloudformation-params.json "${S3_PATH}"
# Upload environment-specific parameters if they exist
if [ -f cloudformation-params.dev.json ]; then
aws s3 cp cloudformation-params.dev.json "${S3_PATH}"
fi
if [ -f cloudformation-params.prd.json ]; then
aws s3 cp cloudformation-params.prd.json "${S3_PATH}"
fi
# Upload any build artifacts as needed
# Example: aws s3 cp build/function.zip "${S3_PATH}"
- name: Deployment info
run: |
echo "✓ Deployed version: ${{ steps.version.outputs.version }}"
echo "✓ S3 path: s3://${{ secrets.S3_ARTIFACT_BUCKET }}/${{ steps.version.outputs.s3_prefix }}/"
echo "✓ Environment: dev"
Replace placeholders:
<repo-owner>/<repo-name>: Full repository name from Git (e.g.,mycompany/my-app)<build-command>: Detected or user-provided build command
Note: The workflow uses ${{ secrets.AWS_REGION }} which will be configured as a GitHub secret.
Step 5: Update CloudFormation Template
If cloudformation.template exists, read it and check for required parameters.
If any of these parameters are missing, add them:
Parameters:
Env:
Type: String
Default: dev
Description: Environment name (dev, staging, prd)
AllowedValues:
- dev
- staging
- prd
Version:
Type: String
Description: Build version in format {build_number}.{commit_hash}
S3Bucket:
Type: String
Description: Artifacts bucket name
S3Prefix:
Type: String
Description: S3 path to artifacts in format {repo}/{branch}/{version}
If cloudformation.template doesn't exist, create a simple example:
AWSTemplateFormatVersion: '2010-09-09'
Description: Deployment for <repo-name>
Parameters:
Env:
Type: String
Default: dev
Description: Environment name (dev, stg, prd, etc)
Version:
Type: String
Description: Build version in format {build_number}.{commit_hash}
S3Bucket:
Type: String
Description: Artifacts bucket name
S3Prefix:
Type: String
Description: S3 path to artifacts in format {repo}/{branch}/{version}
Resources:
# TODO: Add your infrastructure resources here
# Example:
# MyBucket:
# Type: AWS::S3::Bucket
# Properties:
# BucketName: !Sub '${Env}-<repo-name>-${AWS::AccountId}'
# Tags:
# - Key: Version
# Value: !Ref Version
# - Key: Environment
# Value: !Ref Env
Common usage patterns:
Lambda Function
MyFunction:
Type: AWS::Lambda::Function
Properties:
FunctionName: !Sub '${Env}-my-function'
Code:
S3Bucket: !Ref S3Bucket
S3Key: !Sub '${S3Prefix}/function.zip'
Environment:
Variables:
VERSION: !Ref Version
ECS Task Definition
TaskDefinition:
Type: AWS::ECS::TaskDefinition
Properties:
Family: !Sub '${Env}-my-service'
ContainerDefinitions:
- Name: app
Image: !Sub '${AWS::AccountId}.dkr.ecr.${AWS::Region}.amazonaws.com/my-app:${Version}'
Step 6: Create Parameter Files
Base Parameters File
Create cloudformation-params.json as a template:
{
"Env": "dev",
"Version": "1.000000",
"S3Bucket": "<bucket-name>",
"S3Prefix": "<repo-owner>/<repo-name>/main/1.000000"
}
Replace:
<bucket-name>: S3 bucket name provided by user<repo-owner>/<repo-name>: Full repo name from Git
Note: The GitHub workflow dynamically generates this file during deployment, but this template file:
- Shows the structure for reference
- Can be used for local testing (users can modify values manually)
Environment-Specific Parameter Files
Environment-specific parameter files allow you to override or add parameters per environment. They merge with the base cloudformation-params.json.
How they work:
- AWS Deployer loads
cloudformation-params.jsonfirst - Then merges
cloudformation-params.<env>.jsonif it exists - Environment-specific values override base values
When to use them:
- Different instance sizes per environment (t3.micro in dev, m5.large in prod)
- Feature flags (debug mode enabled only in dev)
- Environment-specific endpoints or domains
- Different scaling parameters
Create cloudformation-params.dev.json:
{
}
Create cloudformation-params.prd.json:
{
}
Example with actual parameters:
If your CloudFormation template has:
Parameters:
InstanceType:
Type: String
Default: t3.micro
EnableDebugMode:
Type: String
Default: "false"
You might use:
cloudformation-params.dev.json:
{
"InstanceType": "t3.micro",
"EnableDebugMode": "true"
}
cloudformation-params.prd.json:
{
"InstanceType": "m5.large",
"EnableDebugMode": "false"
}
Step 7: Explain GitHub OIDC Setup
After creating all files, explain to the user how to set up GitHub OIDC authentication and configure the required secrets. Make sure to clearly communicate that they need to:
- Run the
aws-deployer setup-githubCLI command (you do NOT run it for them) - Manually add the
AWS_REGIONGitHub secret - Verify all three secrets are configured:
AWS_ROLE_ARN,S3_ARTIFACT_BUCKET,AWS_REGION
Provide the following instructions:
Next: Configure GitHub OIDC Authentication
Your repository is now configured for AWS Deployer! Before the GitHub Actions workflow can run, you need to set up GitHub OIDC authentication and configure GitHub secrets.
Required GitHub Secrets: Your workflow requires THREE GitHub secrets to be configured:
AWS_ROLE_ARN- IAM role for GitHub OIDC authenticationS3_ARTIFACT_BUCKET- S3 bucket for storing deployment artifactsAWS_REGION- AWS region where the deployer is running
Step 1: Set Up GitHub OIDC with AWS
You need to run the aws-deployer setup-github CLI command to configure GitHub OIDC authentication.
Prerequisites:
-
GitHub Personal Access Token (PAT) stored in AWS Secrets Manager:
# Create a PAT at https://github.com/settings/tokens with 'repo' scope # Then store it in Secrets Manager: aws secretsmanager create-secret \ --name github/pat-token \ --secret-string '{"github_pat":"ghp_xxxxxxxxxxxxx"}' \ --region <region> -
AWS Deployer CLI built and available:
# Build the CLI if not already available cd /path/to/aws-deployer make build-cli
Run the setup command:
aws-deployer setup-github \
--role-name github-actions-<short-repo-name> \
--repo <repo-owner>/<repo-name> \
--bucket <bucket-name> \
--github-token-secret github/pat-token \
--region <region>
For your repository:
aws-deployer setup-github \
--role-name github-actions-<short-repo-name> \
--repo <full-repo-name> \
--bucket <bucket-name> \
--github-token-secret github/pat-token \
--region <region>
What this command does:
- Creates GitHub OIDC provider in AWS (if it doesn't exist)
- Creates IAM role
github-actions-<repo-name>with S3 upload permissions - Adds
AWS_ROLE_ARNsecret to your GitHub repository - Adds
S3_ARTIFACT_BUCKETsecret to your GitHub repository
Step 2: Add Required GitHub Secrets
After running the aws-deployer setup-github command, verify that these GitHub secrets exist. You may need to add AWS_REGION manually:
Required secrets:
AWS_ROLE_ARN- Created byaws-deployer setup-githubcommandS3_ARTIFACT_BUCKET- Created byaws-deployer setup-githubcommandAWS_REGION- Must be added manually
To add AWS_REGION manually:
- Go to
https://github.com/<owner>/<repo>/settings/secrets/actions - Click "New repository secret"
- Name:
AWS_REGION - Value:
<region>(e.g.,us-west-2) - Click "Add secret"
Verify all secrets exist:
Go to https://github.com/<owner>/<repo>/settings/secrets/actions and confirm:
- ✓
AWS_ROLE_ARN- IAM role ARN for GitHub OIDC - ✓
S3_ARTIFACT_BUCKET- S3 bucket name (e.g.,<bucket-name>) - ✓
AWS_REGION- AWS region (e.g.,<region>)
Step 3: Test the Workflow
Once all secrets are configured:
- Push a commit or click "Run workflow" in the Actions tab
- Monitor the deployment in GitHub Actions
- Check AWS Step Functions for execution status
Monitoring Deployments
After pushing code, monitor the deployment:
GitHub Actions
- Check the Actions tab:
https://github.com/<owner>/<repo>/actions - Review workflow logs for any errors
AWS Step Functions
# Get state machine ARN
STATE_MACHINE_ARN=$(aws ssm get-parameter \
--name "/<env>/aws-deployer/state-machine-arn" \
--query 'Parameter.Value' \
--output text)
# List recent executions
aws stepfunctions list-executions \
--state-machine-arn "${STATE_MACHINE_ARN}" \
--max-results 10
Lambda Logs
# S3 trigger logs
aws logs tail /aws/lambda/<env>-aws-deployer-s3-trigger --follow
# Build trigger logs
aws logs tail /aws/lambda/<env>-aws-deployer-trigger-build --follow
DynamoDB Build Records
# Check build history
aws dynamodb query \
--table-name <env>-aws-deployer--builds \
--key-condition-expression "pk = :pk" \
--expression-attribute-values '{":pk":{"S":"<repo-name>/dev"}}'
Troubleshooting
"Error: Could not assume role with OIDC"
Cause: GitHub OIDC not configured or secrets missing
Fix:
- Run the
aws-deployer setup-githubcommand shown above - Verify all required secrets exist in GitHub repository settings:
AWS_ROLE_ARNS3_ARTIFACT_BUCKETAWS_REGION
- Check IAM role trust policy allows your repository
aws iam get-role --role-name github-actions-<repo-name> \
--query 'Role.AssumeRolePolicyDocument'
"S3_ARTIFACT_BUCKET secret not found"
Cause: GitHub secret not configured
Fix: Run the aws-deployer setup-github command to create the secret, or manually add it:
- Go to
https://github.com/<owner>/<repo>/settings/secrets/actions - Click "New repository secret"
- Name:
S3_ARTIFACT_BUCKET - Value: Your S3 bucket name
S3 Upload Succeeds but Deployment Doesn't Start
Cause: S3 trigger Lambda not configured or wrong S3 path
Fix:
# Check S3 files were uploaded
aws s3 ls s3://<bucket-name>/<repo-owner>/<repo-name>/ --recursive
# Check S3 trigger Lambda logs
aws logs tail /aws/lambda/<env>-aws-deployer-s3-trigger --since 10m
CloudFormation Stack Fails
Check stack events:
# List stacks for this repo
aws cloudformation describe-stacks \
--query 'Stacks[?contains(StackName, `<repo-name>`)].{Name:StackName, Status:StackStatus}'
# Get detailed errors
aws cloudformation describe-stack-events \
--stack-name <env>-<repo-name> \
--max-items 20
Files Created
After running this skill, you should have:
-
.github/workflows/deploy.yml- GitHub Actions workflow -
cloudformation.template- CloudFormation template with required parameters -
cloudformation-params.json- Base parameters template -
cloudformation-params.dev.json- Dev environment parameters (empty initially) -
cloudformation-params.prd.json- Production environment parameters (empty initially)
Summary
This skill configured your repository for AWS Deployer by:
- ✓ Detecting repository information from Git
- ✓ Getting S3 bucket and AWS region (from environment variables or by asking)
- ✓ Creating GitHub Actions workflow
- ✓ Ensuring CloudFormation template has required parameters
- ✓ Creating parameter files for local testing
Next steps:
- Run
aws-deployer setup-githubcommand to configure GitHub OIDC - Add
AWS_REGIONGitHub secret manually - Verify all three secrets exist:
AWS_ROLE_ARN,S3_ARTIFACT_BUCKET,AWS_REGION - Push a commit to trigger your first deployment
- Add your infrastructure resources to
cloudformation.template - Add environment-specific parameters if needed
For infrastructure setup, use the setup-deployer skill.
AWS Deployer Setup and Configuration
This skill guides users through the complete setup and configuration of AWS Deployer (github.com/savaki/aws-deployer), a serverless CloudFormation deployment automation system that orchestrates infrastructure deployments using AWS Step Functions, Lambda, and DynamoDB.
What is AWS Deployer?
AWS Deployer is a serverless system that automates CloudFormation deployments across single or multiple AWS accounts. It:
- Monitors S3 for CloudFormation templates and automatically deploys them
- Uses Step Functions to orchestrate complex multi-stage deployments
- Supports promotion workflows (dev → staging → production)
- Integrates with GitHub Actions via OIDC for secure CI/CD
- Provides optional web UI for deployment monitoring
Your Task
When this skill is invoked, you are helping the user set up AWS Deployer from scratch. Your role is to interactively guide them through the complete setup process:
- Gather requirements - Ask about their deployment needs (single vs multi-account, custom domain, etc.)
- Verify prerequisites - Ensure they have AWS CLI, Go, S3 bucket, and necessary permissions
- Build the project - Guide them to compile all Lambda functions
- Deploy infrastructure - Help construct and execute the CloudFormation deployment command
- Configure targets - Set up deployment targets for multi-account mode (if applicable)
- Setup GitHub integration - Configure OIDC roles and secrets for CI/CD
- Verify setup - Confirm all infrastructure is deployed correctly
Important: This is an interactive, step-by-step process. Don't just provide all the instructions at once. Ask questions, wait for responses, verify each step is complete before moving to the next one, and adapt to the user's specific needs.
Prerequisites Check
Before starting, verify the user has:
- AWS CLI installed and configured with appropriate credentials
- Go 1.24+ installed
- An S3 bucket for artifacts (ask for bucket name or help create one)
- IAM permissions to create:
- Lambda functions
- Step Functions state machines
- DynamoDB tables
- IAM roles and policies
- CloudFormation stacks
- For multi-account: IAM roles in target accounts
- For custom domain: Route 53 hosted zone in the deployment account managing the domain's nameservers
Setup Process
1. Gather Requirements
Ask the user:
- Environment name: What environment are they setting up? (dev, stg, prd)
- Deployment mode: Single-account or multi-account deployments?
- S3 bucket: Do they have an existing artifacts bucket, or should one be created?
- Custom domain: Do they want to set up a custom domain for the API Gateway? (optional)
- Important: The AWS account must manage the domain's nameservers via Route 53
- The Route 53 hosted zone must already exist
- They'll need the Hosted Zone ID
- Authorization: Should access be restricted to specific email addresses? (optional)
- AWS Region: Which region to deploy to? (default: us-east-1)
2. Build the Project
Guide the user to build all Lambda functions:
make build
Verify the build succeeded by checking for .zip files in the build/ directory.
3. Set Up Custom Domain Prerequisites (Optional)
If the user wants a custom domain, guide them through these prerequisites:
Verify Route 53 Hosted Zone:
# List hosted zones to find the Zone ID
aws route53 list-hosted-zones-by-name --dns-name example.com
# Verify nameservers are properly configured
aws route53 get-hosted-zone --id <ZONE_ID>
Important: The domain's nameservers at the registrar must point to the AWS Route 53 nameservers shown in the hosted zone.
Create/Verify ACM Certificate:
# Request a certificate (if not already exists)
aws acm request-certificate \
--domain-name deployer.example.com \
--validation-method DNS \
--region <region>
# List certificates to find the ARN
aws acm list-certificates --region <region>
# Get certificate details and validation records
aws acm describe-certificate --certificate-arn <CERTIFICATE_ARN> --region <region>
Important:
- The ACM certificate must be in the same region as the deployment
- DNS validation records must be added to Route 53
- Wait for certificate status to be
ISSUEDbefore deploying
Verify Certificate is Issued:
aws acm describe-certificate \
--certificate-arn <CERTIFICATE_ARN> \
--region <region> \
--query 'Certificate.Status' \
--output text
Should return ISSUED before proceeding.
4. Deploy Infrastructure
Help construct the deployment command based on their requirements:
Basic deployment:
ENV=dev S3_BUCKET=<bucket-name> make deploy
With custom domain:
ENV=prd \
S3_BUCKET=<bucket-name> \
ZONE_ID=Z1234567890ABC \
DOMAIN_NAME=deployer.example.com \
CERTIFICATE_ARN=arn:aws:acm:us-east-1:123456789012:certificate/abc-123 \
make deploy
With all optional parameters:
ENV=prd \
S3_BUCKET=<bucket-name> \
DEPLOYMENT_MODE=multi \
ALLOWED_EMAIL=admin@example.com \
ZONE_ID=Z1234567890ABC \
DOMAIN_NAME=deployer.example.com \
CERTIFICATE_ARN=arn:aws:acm:... \
make deploy
Monitor the CloudFormation deployment and help troubleshoot any errors.
5. Verify Infrastructure
After deployment, verify the CloudFormation stack deployed successfully:
# Check CloudFormation stack status
aws cloudformation describe-stacks \
--stack-name <env>-aws-deployer \
--query 'Stacks[0].StackStatus' \
--output text
Should return CREATE_COMPLETE or UPDATE_COMPLETE.
If the stack deployed successfully, all resources (Lambda functions, DynamoDB tables, Step Functions, Parameter Store) are configured automatically.
6. Set Up Deployment Targets (Multi-Account Mode Only)
If using multi-account mode, configure deployment targets:
Build the CLI:
make build-cli
Configure initial environment:
aws-deployer targets config \
--env <env> \
--default \
--initial-env dev
Set up deployment targets for each environment:
Note: --downstream-env specifies where this environment promotes TO (the next environment in the pipeline).
For dev (promotes to staging):
aws-deployer targets set \
--env <env> \
--target-env dev \
--default \
--accounts "123456789012" \
--regions "us-east-1" \
--downstream-env "staging"
For staging (promotes to prd):
aws-deployer targets set \
--env <env> \
--target-env staging \
--default \
--accounts "123456789012" \
--regions "us-east-1,us-west-2" \
--downstream-env "prd"
For production (final environment, no promotion):
aws-deployer targets set \
--env <env> \
--target-env prd \
--default \
--accounts "123456789012,987654321098" \
--regions "us-east-1,us-west-2,eu-west-1"
Verify target configuration:
aws-deployer targets list --env <env>
7. Set Up GitHub Repository for CI/CD
For each repository that will use AWS Deployer:
Prerequisites:
- GitHub Personal Access Token (PAT) with
reposcope stored in AWS Secrets Manager - Your AWS credentials must have
secretsmanager:GetSecretValuepermission for the GitHub PAT secret
Create the GitHub PAT secret (if not already exists):
# Create the secret with your GitHub PAT
aws secretsmanager create-secret \
--name github/pat-token \
--secret-string '{"github_pat":"ghp_xxxxxxxxxxxxx"}'
Create GitHub OIDC role and secrets:
aws-deployer setup-github \
--role-name github-actions-<repo-name> \
--repo owner/repository-name \
--bucket <artifacts-bucket> \
--github-token-secret github/pat-token
This creates:
- IAM OIDC provider for GitHub
- IAM role with S3 access scoped to the repository path
- GitHub repository secret with the role ARN (using the PAT from Secrets Manager)
Provide the user with a sample GitHub Actions workflow:
name: Deploy to AWS
on: [push]
permissions:
id-token: write
contents: read
jobs:
deploy:
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- name: Configure AWS Credentials
uses: aws-actions/configure-aws-credentials@v4
with:
role-to-assume: ${{ secrets.AWS_ROLE_ARN }}
aws-region: us-east-1
- name: Build and package
run: |
# Build your application and create CloudFormation template
# Generate cloudformation-params.json with standard parameters
REPO="<repo>"
BRANCH="${{ github.ref_name }}"
SHA_SHORT="${{ github.sha }}"
SHA_SHORT="${SHA_SHORT:0:6}"
VERSION="${{ github.run_number }}.${SHA_SHORT}"
S3_PREFIX="${REPO}/${BRANCH}/${VERSION}"
cat > cloudformation-params.json <<EOF
{
"Env": "dev",
"Version": "${VERSION}",
"S3Bucket": "<bucket>",
"S3Prefix": "${S3_PREFIX}"
}
EOF
- name: Upload to S3
run: |
REPO="<repo>"
BRANCH="${{ github.ref_name }}"
SHA_SHORT="${{ github.sha }}"
SHA_SHORT="${SHA_SHORT:0:6}"
VERSION="${{ github.run_number }}.${SHA_SHORT}"
S3_PATH="s3://<bucket>/${REPO}/${BRANCH}/${VERSION}/"
aws s3 cp cloudformation.template "${S3_PATH}"
aws s3 cp cloudformation-params.json "${S3_PATH}"
Important: Your CloudFormation templates should always expect these standard parameters:
Env- Environment name (dev, staging, prd)Version- Build version in format{build_number}.{commit_hash}S3Bucket- Artifacts bucket nameS3Prefix- S3 path to artifacts in format{repo}/{branch}/{version}
8. Multi-Account IAM Setup (Multi-Account Mode Only)
For multi-account deployments, set up IAM roles in target accounts:
aws-deployer setup-aws \
--account-id <target-account-id> \
--admin-account-id <deployer-account-id> \
--role-name AWSCloudFormationStackSetExecutionRole
This creates the execution role that CloudFormation StackSets need in each target account.
Troubleshooting
Help the user diagnose common issues:
S3 Trigger Not Working
- Verify S3 bucket notification is configured correctly
- Check Lambda permissions to be invoked by S3
- Review S3 trigger Lambda logs
Step Function Not Starting
- Verify
state-machine-arnin Parameter Store - Check DynamoDB stream configuration
- Review trigger-build Lambda logs
Multi-Account Deployment Failures
- Verify target accounts are configured in DynamoDB
- Check IAM execution roles exist in target accounts
- Verify StackSet administration role has permissions
- Review check-stackset-status Lambda logs for detailed errors
Lock Acquisition Timeouts
- Check if another deployment is running
- Verify lock TTL hasn't expired prematurely
- Review acquire-lock Lambda logs
Parameter Store Access Issues
- Verify Lambda IAM roles have
ssm:GetParameter*permissions - Check parameter paths match the environment name
- Consider using
DISABLE_SSM=truefor local development
Completion Verification
Verify the setup is complete by checking the CloudFormation stack status:
# Check stack status
aws cloudformation describe-stacks \
--stack-name <env>-aws-deployer \
--query 'Stacks[0].StackStatus' \
--output text
Should return CREATE_COMPLETE or UPDATE_COMPLETE.
If the CloudFormation stack deployed successfully, all infrastructure resources are automatically created:
- Lambda functions with correct IAM permissions
- DynamoDB tables (builds, targets, deployments, locks)
- Step Function state machines (single and multi-account)
- S3 bucket notification configuration
- Parameter Store values
- API Gateway and custom domain (if configured)
Additional verification for multi-account setups:
- Deployment targets configured (if using multi-account mode)
- Target account IAM roles created (if using multi-account mode)
Additional verification for GitHub integration:
- GitHub repository configured with OIDC role
- Test deployment completed successfully
Next Steps
After setup is complete, guide the user to:
- Use the
deployskill to test their first deployment - Review CLAUDE.md for development workflow
- Read MULTI.md for multi-account deployment details (if applicable)
- Explore DEPLOYMENT_TARGETS.md for advanced target configuration
- Set up monitoring and alerting for Step Functions
- Configure additional repositories for automated deployments
Reference Commands
Keep these handy for the user:
# View build history
aws dynamodb query \
--table-name <env>-aws-deployer--builds \
--key-condition-expression "pk = :pk" \
--expression-attribute-values '{":pk":{"S":"<repo>/<env>"}}'
# Update Parameter Store
aws ssm put-parameter \
--name "/<env>/aws-deployer/<param>" \
--value "<value>" \
--overwrite
# View Step Function executions
aws stepfunctions list-executions \
--state-machine-arn <arn> \
--max-results 10
# Tail Lambda logs
aws logs tail /aws/lambda/<env>-aws-deployer-<function> --follow
# Update Lambda code
make update-lambda-code
# Full redeployment
make clean-version && make deploy
Sponsors
Details
- Marketplace
- savaki/savaki-marketplace
- Source
- plugins/aws-deployer
- Author
- Matt Ho
- Category
- Community