alexnederlof / revs3

A super simple reverse proxy for S3. Meant to host websites on an internal network backed by an S3 bucket.

Reverse S3

A super simple reverse proxy for S3 written in Rust. Can be used as a web server with S3 as the backend.

This project was created because S3 Buckets cannot be configured to function as a website on an internal network. If they are configured as a website, it is always public. This project aims to solve that problem.

The more popular option is NGinx with S3 backend but its Auth module cannot deal with AWS_WEB_IDENTITY_TOKEN_FILE as an auth mechanism and that's what's needed if you run this on Kubernetes.

Configuration

Besides the config below, you also need some form of AWS Authentication set through environment variables. See the official docs for specifics

Configuration Option	Description	Default	Required
AWS_REGION	The AWS region where your S3 bucket is located.		Yes
S3_BUCKET	The name of your S3 bucket.		Yes
LOG_LEVEL	The log level for the reverse proxy server.	info	No
KEY_PREFIX	A prefix you want to use, in case your files are in a subfolder of S3		No

See example configuration in .env. This file is excluded from the Docker build so needs to be set server side.

Besides this, the service needs a least the following permissions to be able to function:

{
  "Version": "2012-10-17",
  "Statement": [
    {
      "Sid": "AllowReadAccessToBucket",
      "Effect": "Allow",
      "Action": ["s3:GetObject", "s3:ListBucket"],
      "Resource": [
        "arn:aws:s3:::your-bucket-name/*",
        "arn:aws:s3:::your-bucket-name"
      ]
    }
  ]
}

Caching

The server does no caching itself. It relies on the E-Tag headers to enable browser caching. Further more, it will relay any expire headers set by S3 to the browser.

Developing

Using docker-compose up -d will bootstrap localstack to mimic S3, and will put files from test_data in that bucket.

The test data assumes a bucket prefix called "reports". This can be omitted.