TL;DR;

This Terraform script lets you deploy a Lambda function on CloudFront that will compress your images on demand. You can use the Lambda function on an existing distribution, or create a separate one to serve your files stored in S3. To do so, prefix all your images with https://images.your_domain.com/ to compress them on the fly, for example:

https://images.your_domain.com/image_in_my_bucket.png?width=200&webp=1

... will compress then cache the origin image to a 200px-wide WebP image with a preserved aspect ratio.

Great to generate responsive images tags from one single image.

Image compression "at the edge"

Terraform module which lets you supercharge your website by automatically compressing your images on CloudFront, using Lambda@Edge

Based on https://github.com/lovell/sharp for faster compression than ImageMagick

These types of resources are supported:

• CloudFront Distribution
• Lambda@Edge

How does it work - What does it solve?

If your website or web application displays user-generated images, you need to resize and compress them to keep your website as fast as possible. A classic way to handle this is to have a Lambda function triggered on the receiving S3 bucket that would compress them in place, possibly in various sizes. However it does not work if images are hosted by a third-party partner and when your front-end is changing and you want to offer different sizes later down the road.

This CloudFront-based Lambda function will let you update the image size at any time.

For an original image of 527KB:

• width 400px, size 26KB

• width 300px, size 16.7KB

• width 200px, size 8.4KB

• width 100px, size 3.1KB

Usage

Clone this Terraform files and replace variables in terraform.tfvars by your own values.

Edit: You need to replace node_modules/sharp by a version compiled on a compatible operating system, so the build is compatible with Lambda execution environment. I will document this later, in the mean time, know that you can build it on a ami-489f8e2c:

docker-machine create -d amazonec2 --amazonec2-access-key YOUR_ACCESS_KEY --amazonec2-secret-key YOUR_SECRET_KEY --amazonec2-instance-type t2.small --amazonec2-ami ami-489f8e2c --amazonec2-region eu-west-2 buildmachine

cd lambda/code/origin_response

npm install

Run terraform apply and enjoy

Authors

Module managed by Lionel Martin

License

Apache 2 Licensed. See LICENSE for full details