How to Build a Cloud‑Native Microservices PaaS with Spring Cloud Netflix

The goal of the PaaS cloud platform is to give developers a streamlined workflow for rapid development, deployment, operation, and continuous integration by providing infrastructure, middleware, data services, and servers; developers only need to commit business code and configure the system, which then automatically builds and deploys the application.

Implementing microservices requires substantial effort, so the article recommends mature open‑source frameworks—Netflix, Spring Cloud, and Dubbo—focusing on Spring Cloud Netflix, which builds on Spring Boot and supplies essential components for microservice development.

Service Gateway (GateWay)

The gateway sits between external clients (browsers, mobile devices) and internal services. It is deployed in a cluster behind Amazon EC2 instances with Elastic Load Balancing (ELB). SSL termination is handled by Nginx. The gateway provides dynamic routing, rate limiting, authentication, monitoring, and access logging. Spring Cloud Netflix’s Zuul is used, leveraging a series of customizable filters to implement these functions.

Service Registration and Discovery

Service registration and discovery are handled by Eureka. Each microservice registers with the Eureka server and sends heartbeats (default every 30 seconds). The server marks a service as dead after missing three heartbeats (default 90 seconds) unless self‑protection mode is active. Eureka nodes are clustered and synchronize registration data via DNS‑based discovery.

Microservice Deployment

Microservices run as Docker containers. Docker images (including base JDK images) are built for each service and stored in a private image repository. Containers are managed as a cluster using Docker Swarm, allowing multiple services per container and simplifying scaling and updates.

Service Fault Tolerance

Hystrix (a Netflix open‑source library) provides fault tolerance through circuit breaking, thread isolation, fallback mechanisms, and rate limiting. Circuit breakers trigger when error rates exceed 50 %, entering a half‑open state after a cooldown period. Thread pools isolate services, preventing a failure in one service from exhausting resources of others. Fallback logic is implemented by extending HystrixCommand and overriding getFallback(). Rate limiting caps concurrent requests to protect downstream services.

Dynamic Configuration Center

Spring Cloud Config Server supplies a dynamic configuration center. Configuration files are kept in a private Git repository. When developers push changes, a server‑side hook notifies Config Server via a message queue, prompting it to refresh the affected configurations. This enables runtime updates without rebuilding or redeploying services.

In practice, the architecture combines Zuul (gateway), Eureka (service registry), Hystrix (fault tolerance), and Ribbon (client‑side load balancing) along with additional logging and messaging components as needed.