Spring-boot with Bucket4j and Hazelcast Demo

The aim of this demo is to demonstrate rate-limiting by using Bucket4j and the In-Memory Data Grid Hazlecast.

To do that, we will run two instances of the same application, they share the same bucket which is stored on Hazelcast. When a request from the same Ip hit any of the two instances the shared bucket is decremented.

The result is immediately visible to any instance.

Demo

1. Run the first instance : mvn spring-boot:run -Dserver.port=8081

2. Run the second instance : mvn spring-boot:run -Dserver.port=8082

3. You can also run a third instance if you want.

Testing

When calling any of the two instances, say we run the following command curl -s -D - localhost:8081, the response looks like :

 HTTP/1.1 200
 X-Rate-Limit-Remaining: 9
 Content-Type: text/plain;charset=UTF-8
 Content-Length: 13
 Date: Thu, 29 Jun 2017 21:25:00 GMT
 
 Hello World !

Notice the X-Rate-Limit-Remaining that indicates the remaining requests.

In this sample we configured the bucket with a capacity of 10 requests every minute as the following code shows.

Bucket4j.configurationBuilder()
               .addLimit(Bandwidth.simple(10, Duration.ofMinutes(1)))
               .buildConfiguration();

Run alternatively the following commands :

1. Request the instance one : curl -s -D - localhost:8081

2. Request the instance two : curl -s -D - localhost:8082

You will see the value of X-Rate-Limit-Remaining header decrementing until it equals 0 the subsequent request gives a 429 HTTP status code along with message too many requests.

 HTTP/1.1 429
 Content-Type: text/plain;charset=ISO-8859-1
 Content-Length: 17
 Date: Thu, 29 Jun 2017 21:36:11 GMT
 
 Too many requests

Note :

It is possible to not see the X-Rate-Limit-Remaining decrementing or on contrary incrementing, especially if you take a long time to execute the queries, this is because bucket4j regenerates tokens immediately (for more detailed information read the documentation of Refill.java )