This service is a simple reverse proxy that can be configured to route requests to different services based on the request path. It also has a rate limiter that can be configured to limit the number of requests that can be made to a given route.

The reverse proxy implementation can be found in cmd/proxy_server/main.go, there are also two example servers that can be used to test the proxy in cmd/server_one/main.go and cmd/server_two/main.go.

• Docker
• Go 1.22.*
• go-task

1. Spin up the docker containers using docker-compose. This will start the following services:

   • proxy_server: The reverse proxy server, runs on :6666
   • server_one: A basic server that returns "Hello", runs on localhost:8080/hello
   • server_two: A basic server that returns "Goodbye", runs on localhost:9090/goodbye
   • redis: A redis server used to store rate limiting information
   • token_bucket_refiller: A service that refills the token buckets for the rate limiter

   $ docker-compose up

2. Make requests to the proxy, see route definitions in RouteDefinitions.yml:

   curl localhost:6666/api/hello
   curl localhost:6666/api/goodbye

The proxy is configured by a YAML file that is read in at startup. The configuration file is located at ./RouteDefinitions.yml.

The configuration file is a list of routes to be mounted by the proxy, and a Target field that specifies the address of the service that the proxy should route requests to. The rate_limit field specifies the rate limit for the route in requests per minute.

  - path: "/api/goodbye"  # Route to be mounted by the proxy
    method: "GET"         # HTTP assumed to be same on both sides
    target:               # Target configuration
      host: "server_one"
      port: ":9090"
      path: "/goodbye"
    rate_limit:
      requests_per_minute: 1200 # Number of requests per minute -- ensure divisible by 60 for exact limiting

• Add a boat load of tests including a long running integration test that hits the proxy with a bunch of requests.
• Implement an additional rate limiter that tracks a user's requests across all routes (global limit).
• Implement different rate limiting strategies that can be configured via the proxy config file.
• Look at mTLS/Ingress/Firewalls
• Dockerise and test with minikube (Services/Ingresses/NetworkPolicies)