Microservice Interview - 4 : Circuit Breaker

Hi Friends,

In this blog post, I'm explaining the Circuit Breaker pattern that will cover all interview questions on this.


It is very important to prepare system in case of partial failure, especially for a microservice-based architectures, where there are many applications running in separate processes.

A single request from the client point of view might be forwarded through many different services and it is possible that one of thee services is down because of a failure, maintenance or just might be overloaded , which causes an extremely slow response to client requests coming into the system.

There are several best practices for dealing with failures and errors:

• The first practice recommends that we should always set network connect and read timeouts to avoid waiting too long for the response.
• The second approach is about limiting the number of accepted requests if a service fails or responses take too long. In this case, there is no sense in sending additional requests by the client.

The last two patterns are closely connected to each other. Here I'm thinking about the circuit breaker pattern and fallback.

Few points about the Circuit breaker diagram:

• Initially the circuit is closed. Means the service is working fine , accepting requests and returning response.

• When few requests get failed, circuit gets tripped and enter into Open state and doesn't allow any further request to go to that non-responding service.

• In between, after some timeout , circuit breaker allows few requests to the non-responding service to check the status of that service. This state is called Half-open state as it allows only few requests to go further. If service responds well, then circuit gets into Closed state , else it will get into Open state again.

Circuit Breaker Pattern:

The major assumption of this approach relies on monitoring successful and failed requests. If too many requests fail or services take too long to respond, the configured circuit breaker is tripped and all further requests are rejected.

The most  popular implementation of Circuit Breaker pattern is available in Netflix Hystrix which is used by many java frameworks like Spring Cloud or Apache Camel.

Implementation of a circuit breaker with Spring Cloud Netflix is quite simple. 

In the main class, it can be enabled with one annotation:

@SpringBootApplication

@EnableFeignClients

@EnableCircuitBreaker

public class Application{

    public static void main(String[] args){

        SpringApplication.run(Application.class, args);

    }

}

And also need to add dependency :

In gradle:  Use :

compile group: 'org.springframework.cloud', name: 'spring-cloud-starter-hystrix', version: '1.4.7.RELEASE'

For fallback mechanism, we can use FeignClient like:

Our service A is going to interact with service B , but circuit breaker is tripped and requests to service B are rejected.

So , we need to call any fallback method for providing the response.

A.java:

@FeignClient(name = "A",  url = " URL or endpoint of service B" , fallback = A.Fallback.class)

public interface A{

    @GetMapping(value = "/{id}?app_id={app_id}")

    ResponseEntity<String> getInfo(@PathVariable("id") String id);

    class Fallback implements A{

        @Override

        public ResponseEntity<String> getInfo(String id){

           // Send back any info.

        }

    }

}

A pictorial diagram is like:

That's all for this post.

Hope this post helps everyone in their interviews.

Thanks for reading!!