Microservice is a service-oriented architecture where an application is deployed as a collection of loosely-couple services. The goal is to make each service independent, fine-grained, scalable and flexible which allows faster testing and release.
applicant-services - Dummy service that returns a list of applicant names.
@GetMapping("/applicants-by-job")
public List<String> getApplicantsByJob() {
log.debug("port={} get applicants by job", port);
return Arrays.asList("Steve", "Bill", "Linus");
}job-services - Dummy service that returns a job title with a list of applicant names.
@GetMapping("/job-with-applicant-profiles")
public ResponseEntity listJobsWithApplicantProfiles() {
log.debug("get job details with applicants");
JobWithApplicantsDto result = new JobWithApplicantsDto();
result.setJob("Java Developer");
result.setApplicants(applicantProxy.getApplicantsByJob());
return ResponseEntity.ok().body(result);
}- naming-server / discovery-server
This server holds all the information about our microservices such as name, IP, and port. This information is used for exchanging a service name to IP address and port.
A simple naming-server typically has:
// pom.xml
<dependency>
<groupId>org.springframework.cloud</groupId>
<artifactId>spring-cloud-starter-netflix-eureka-server</artifactId>
</dependency>
// in application.yml
eureka:
client:
register-with-eureka: false
fetch-registry: false
// in SpringBootApplication annotated class
@EnableEurekaServer- api-gateway
It's a server that provides criteria-driven request routing. It also offers other features such as security, load balancing, logging, monitoring, etc.
API Gateway is not at all different to the naming-server. Thanks to Spring for doing the heavy lifting:
// pom.xml
<dependency>
<groupId>org.springframework.cloud</groupId>
<artifactId>spring-cloud-starter-gateway</artifactId>
</dependency>
// @Configuration annotated class
@Bean
public RouteLocator gatewayRouter(RouteLocatorBuilder builder) {
return builder.routes()
.route(p -> p
.path("/get")
.uri("http://httpbin.org"))
.route(p -> p.path("/applicants/**")
.uri("lb://applicant-services"))
.route(p -> p.path("/jobs/**")
.uri("lb://job-services"))
.build();
}- spring-cloud-starter-openfeign - this library is used for referencing a service in the naming server
To enable this feature, @EnableFeignClients must be annotated to a configuration class.
To call an endpoint from another service, an interface must be created with the same method signature as the method from the other service. In this example, we are importing 2 endpoints from the applicant-service.
@FeignClient(name = "applicant-services")
public interface ApplicantProxy {
@GetMapping("/applicants/applicants-by-job")
List<String> getApplicantsByJob();
@GetMapping("/applicants/top-applicants-by-job")
public List<String> getTopApplicantsByJob();
}- spring-cloud-starter-netflix-eureka-client - this library provide the necessary classes for our services to register to the discovery server.
To register to a naming-server, we need to add this configuration in each of the services:
eureka:
client:
serviceUrl:
defaultZone: http://localhost:8761/eureka // make sure that this address is correct- spring-cloud-starter-sleuth - this library helps us trace our requests across different microservices by automatically adding a unique id to the logs.
As we can see in the logs, it started from api-gateway, pass thru job-services, and then application-services. In the 3 logs, notice the common unique id: 3b4039e47fb602a9. We can use this when tracing a request.
http://localhost:8081/applicants/applicants-by-job http://localhost:8081/applicants/top-applicants-by-job
http://localhost:8080/jobs/job-with-applicant-profiles http://localhost:8080/jobs/job-with-top-applicants
http://localhost:8000/applicants/applicants-by-job http://localhost:8000/applicants/top-applicants-by-job http://localhost:8000/jobs/job-with-applicant-profiles http://localhost:8000/jobs/job-with-top-applicants