Building a Microservice Architecture with Spring Cloud, Docker, and Netflix OSS

The original article demonstrates how to use Spring Boot, Spring Cloud, Docker, and several Netflix OSS projects to build a conceptual microservice architecture, providing a runnable example (PiggyMetrics) that can be started with a single command.

Functional Services

The application is split into three core microservices that can be deployed independently:

Account Service

Handles user input and validation for income/expense records, savings, and account settings.

Statistics Service

Computes statistical parameters and stores time‑series data for each account, normalising values by currency and period.

Notification Service

Stores contact information and notification preferences, aggregates data from other services, and sends e‑mail notifications to subscribed users.

Each microservice has its own database (polyglot persistence) – MongoDB is used in the example.

Inter‑service communication is primarily synchronous REST, with optional asynchronous messaging for create/update operations, leading to eventual consistency.

Infrastructure Services

Configuration Service

Spring Cloud Config provides a centralized configuration repository (native profile, classpath files). Clients fetch shared/notification-service.yml and shared/application.yml at startup.

Example bootstrap.yml for a client:

spring:
  application:
    name: notification-service
  cloud:
    config:
      uri: http://config:8888
      fail-fast: true

Dynamic refresh is possible with @RefreshScope; changes can be applied via a POST to /notifications/refresh or a webhook.

Authorization Service

Provides OAuth2 tokens for resource servers. The example uses password grant for user login and client‑credentials grant for inter‑service calls. Spring Cloud Security supplies annotations such as @PreAuthorize to protect endpoints.

@PreAuthorize("#oauth2.hasScope('server')")
@RequestMapping(value = "accounts/{name}", method = RequestMethod.GET)
public List<DataPoint> getStatisticsByAccountName(@PathVariable String name) {
    return statisticsService.findByAccountName(name);
}

API Gateway

Zuul is used as a single entry point, routing requests to the appropriate backend services and handling cross‑cutting concerns like authentication and canary testing.

zuul:
  routes:
    notification-service:
      path: /notifications/**
      serviceId: notification-service
      stripPrefix: false

Service Discovery

Netflix Eureka registers service instances and enables client‑side load balancing. Services announce themselves on startup and send heartbeats; missing heartbeats cause deregistration.

spring:
  application:
    name: notification-service

Load Balancer, Circuit Breaker, and HTTP Client

Ribbon provides client‑side load balancing, Hystrix implements the circuit‑breaker pattern, and Feign offers a declarative HTTP client that integrates with Ribbon and Hystrix.

@FeignClient(name = "statistics-service")
public interface StatisticsServiceClient {
    @RequestMapping(method = RequestMethod.PUT, value = "/statistics/{accountName}", consumes = MediaType.APPLICATION_JSON_UTF8_VALUE)
    void updateStatistics(@PathVariable("accountName") String accountName, Account account);
}

Monitoring Dashboard

Hystrix metrics are streamed via Spring Cloud Bus to Turbine, and visualised with the Hystrix Dashboard.

Log Analysis

Centralised logging can be achieved with the ELK stack (Elasticsearch, Logstash, Kibana).

Security

Production‑grade security should include HTTPS and encrypted configuration properties using JCE keystores.

Infrastructure Automation

Continuous delivery is demonstrated with Docker Compose and Travis CI, building Docker images for each service and tagging them with Git commit hashes.

How to Run the Whole System

Start eight Spring Boot applications, four MongoDB instances, and RabbitMQ. Ensure at least 4 GB of RAM. The system includes a gateway, Eureka registry, Config Server, Auth Server, and the three functional services.

Prerequisites

Install Docker and Docker Compose.

Set environment variables for passwords (CONFIG_SERVICE_PASSWORD, NOTIFICATION_SERVICE_PASSWORD, etc.).

Production Mode

Pull the latest images from Docker Hub and run docker-compose up -d.

Development Mode

Clone the repositories, build artifacts with Maven, and run

docker-compose -f docker-compose.yml -f docker-compose.dev.yml up -d

to build local images and expose ports.

Important Endpoints

http://localhost:80 – API Gateway

http://localhost:8761 – Eureka Dashboard

http://localhost:9000 – Hystrix Dashboard

http://localhost:8989 – Turbine stream

http://localhost:15672 – RabbitMQ Management

All Spring Boot applications require the Config Server to start; the fail‑fast property and Docker restart policies ensure containers keep trying until the Config Server is available. Service discovery may need a few heartbeats before clients can locate services.