Spring Boot Bucket4j Example for IP-Based Rate-Limiting (3 Alternatives)

An example Spring Boot project which uses Bucket4j rate-limiting library for IP-based rate-limiting. It supports 3 different caching alternatives; Caffeine, Redis and Hazelcast.

Configuration

• For rate-limiting and caching, edit application.yml file to conform your requirements.
• Dependencies in pom.xml file should be set to enable one of Caffeine, Redis or Hazelcast. All dependencies are already included, just uncomment the appropriate section in pom.xml file.

How to Run

Start the application with the command mvn spring-boot:run and test the following endpoints for rate-limiting:

• http://localhost:8080/first - 1 req / 5 sec
• http://localhost:8080/second - 2 req / 5 sec

Distributed Rate Limiting

If distributed rate limiting is enabled (Hazelcast or Redis), then you can start the application on different ports and see the results. For example, you can start 2 instances with the following commands:

    mvn spring-boot:run -D"spring-boot.run.jvmArguments"='-Dserver.port=8080'
    mvn spring-boot:run -D"spring-boot.run.jvmArguments"='-Dserver.port=8081'

Then test with the following endpoints:

• http://localhost:8080/first - 1 req / 5 sec

• http://localhost:8081/first - 1 req / 5 sec

• http://localhost:8080/second - 2 req / 5 sec

• http://localhost:8081/second - 2 req / 5 sec

The path /first will allow only 1 req / 5 sec no matter from which port (8080 or 8081) the page is accessed.

The path /second will allow only 2 req / 5 sec no matter from which port (8080 or 8081) the page is accessed.

Hazelcast:

Hazelcast doesn't need a standalone instance like Redis. Adding the dependencies will be enough to start a Hazelcast cluster.

Redis:

Run the following command to start a Redis instance if you don't have it.

    docker run --name redis -d -p 6379:6379 redis