Spring Cloud for Microservices

Spring Cloud Gateway, Config, service discovery, load balancing, Circuit Breaker, Stream and Kubernetes integration, and which pieces are still needed on modern platforms.

Components and when to use them

Component Purpose On Kubernetes?
Spring Cloud Gateway Reactive (or MVC) API gateway with routes, filters, rate limiting Useful as a code-centric gateway/BFF; alternatives: Envoy Gateway, cloud gateways
Spring Cloud Config Centralized Git-backed configuration Often replaced by ConfigMaps/Secrets, external secret operators, or Spring Cloud Kubernetes
Eureka / Consul discovery Service registry Usually unnecessary: use Kubernetes Services and DNS
Spring Cloud LoadBalancer Client-side load balancing Rarely needed with K8s Services or a mesh
Spring Cloud Circuit Breaker Abstraction over Resilience4j Yes, for application-level fallbacks
Spring Cloud Stream Binder abstraction for Kafka/RabbitMQ with functional style Yes, for event-driven services
Spring Cloud Function Functions portable across serverless platforms Serverless use
Spring Cloud Kubernetes Config and discovery from Kubernetes APIs When apps need dynamic config reload from ConfigMaps

[!VERSION] Spring Cloud Sleuth was replaced by Micrometer Tracing in Spring Boot 3. Netflix Ribbon, Hystrix and Zuul 1 are retired; use Spring Cloud LoadBalancer, Resilience4j and Spring Cloud Gateway.

Gateway route example

spring:
  cloud:
    gateway:
      routes:
        - id: orders
          uri: http://orders-service
          predicates: [ Path=/api/orders/** ]
          filters:
            - StripPrefix=1
            - name: RequestRateLimiter
              args:
                redis-rate-limiter.replenishRate: 50
                redis-rate-limiter.burstCapacity: 100
            - name: CircuitBreaker
              args: { name: orders, fallbackUri: forward:/fallback/orders }

[!VERSION] Spring Cloud Gateway property prefixes changed in recent releases (for example spring.cloud.gateway.server.webflux.*); check the docs for your release train.

Spring Cloud Stream (functional)

@Bean
Function<OrderPlaced, StockReserved> reserveStock(InventoryService inventory) {
    return event -> inventory.reserve(event.orderId(), event.lines());
}
spring.cloud.stream.bindings:
  reserveStock-in-0:  { destination: orders.order-placed.v1, group: inventory }
  reserveStock-out-0: { destination: inventory.stock-reserved.v1 }

Architect-level considerations

[!ARCHITECT] Do not recreate platform features in application libraries. On Kubernetes, rely on the platform for discovery, config and load balancing; keep Spring Cloud for gateway/BFF needs, resilience and messaging abstractions.

Interview questions and answers

Do you still need Eureka?

Not on Kubernetes or most managed container platforms, which provide discovery and load balancing. Eureka remains relevant on VM-based deployments without a platform registry.

How do you refresh configuration without redeploying?

Spring Cloud Kubernetes or Config with @RefreshScope and a refresh trigger, or rolling restarts driven by config changes (often safer and simpler), or a feature flag system for runtime toggles.