AWS Lambda Image Gateway

Image Gateway provides an API endpoint which creates resized versions of images stored inside an S3 bucket.

It uses a combination of S3 bucket, Lambda function, API Gateway and Cloudfront.

Considerations:

• Lifecycle policy should be matched with CloudFront cache, as it is possible for CloudFront to redirect the user to a cached image after this image has been removed. Therefore it is advised that Maximum TTL for S3 origin should be set to 1209600 seconds (14 days).

Lambda function environment:

• Pillow Python module is required
• To create a ready to use package, you have to package the lambda source file (lambda.py) with the Pillow module. To do so, follow instructions here: http://docs.aws.amazon.com/lambda/latest/dg/lambda-python-how-to-create-deployment-package.html

Creating CloudFormation stack:

1. Upload Lambda function ZIP (see above) into an S3 bucket.
2. Upload infrastructure/templates/stack.json into CloudFormation.
3. Fill out parameters:

• BucketName - name of the bucket for storage that will be created
• LambdaSourceBucket - name of the S3 bucket where you uploaded the Lambda function ZIP
• LambdaFileName - name of the Lambda function ZIP file inside the above S3 bucket

1. Ta-dah! You'll find your CloudFront URL inside Outputs in CloudFormation.

To update the Lambda function after creating the stack:

Lambda function is not updated by CloudFormation unless you change the parameters. So you can change the LambdaSourceBucket or LambdaFileName OR update the Lambda manually:

• Get pre-compiled libraries (as above)
• Unzip the archive
• Copy lambda.py into the unpacked directory
• Zip the contents (not the directory) back.
• Upload to AWS Lambda or S3 bucket.