Lambda Image Transformer

A serverless AWS Lambda@Edge function for image transformation using Sharp. Deployed behind CloudFront, it dynamically resizes and converts images stored in S3 via URL query parameters.

How It Works

CloudFront sends an origin request to Lambda@Edge. The Lambda function:

1. Parses the image request from S3
2. Applies transformations via Sharp
3. Returns the optimized image as a base64-encoded response

Features

• Resize images with w (width) and h (height) query parameters
• Convert image formats using format (e.g., webp, jpeg)
• Optimized for AWS Lambda@Edge
• Caches transformed responses via CloudFront

Example Request

https://your-distribution-id.cloudfront.net/uploads/dog.jpg?w=300&format=webp

Returns a 300px-wide WebP version of dog.jpg from S3.

---

Deployment

1. Build the deployment package Use Docker to compile Sharp for Lambda:

docker run -it --rm -v "$PWD":/var/task public.ecr.aws/sam/build-nodejs18.x bash

# Inside container:
npm install
exit
Then:

zip -r lambda-image-transformer.zip .

2. Upload to Lambda (region us-east-1)
Upload the .zip in AWS Lambda Console

Publish a new version

Attach it to CloudFront (Origin Request trigger)

3. IAM Role Requirements
The Lambda execution role must allow:

{
  "Action": "s3:GetObject",
  "Resource": "arn:aws:s3:::your-bucket-name/*"
}
And its trust policy must allow:

"Principal": {
  "Service": ["lambda.amazonaws.com", "edgelambda.amazonaws.com"]
}
Logging
Logs are available in CloudWatch under us-east-1:

/aws/lambda/<FunctionName>
Make sure CloudFront forwards query strings to enable dynamic caching.

📄 License
MIT