Part 10 · 1 chapters · ~8 min

In Production: Spring Cloud and Sidecars

The Java production stack: container images and JVM flags, Spring Cloud (Config, Gateway, discovery, LoadBalancer, Resilience4j), sidecars beside the JVM (Envoy via Istio, Linkerd, Dapr), Kafka with Spring Kafka, the OpenTelemetry Java agent, Vault, and choosing library or sidecar for each concern.

Libraries or sidecars

code

# container image: layered jar, JVM flags for containers
FROM eclipse-temurin:25-jre
COPY build/libs/ledger.jar /app/ledger.jar
ENV JAVA_TOOL_OPTIONS="-XX:MaxRAMPercentage=75 -XX:+UseZGC -javaagent:/otel/opentelemetry-javaagent.jar"
ENTRYPOINT ["java", "-jar", "/app/ledger.jar"]

# Kubernetes pod with an Istio sidecar injected (namespace label istio-injection=enabled)
# the Spring app calls http://limits:8080 in plain HTTP; Envoy adds mTLS, retries and canary routing

concern	Spring Cloud library	sidecar alternative
service discovery	Eureka, Spring Cloud LoadBalancer	Kubernetes DNS + Envoy
resilience	Resilience4j (business-aware fallbacks)	Envoy retries, timeouts, outlier detection
mTLS	manual SSL bundles	automatic (Istio, Linkerd)
configuration	Spring Cloud Config, Spring Cloud Vault	Vault Agent sidecar, External Secrets
pub/sub abstraction	Spring Cloud Stream	Dapr pub/sub

Most mature platforms combine them: sidecars for mTLS, telemetry and coarse traffic policy; libraries where business context matters (fallbacks, idempotent retries).

WHAT JAVA IS PAIRED WITH IN PRODUCTION

Spring Cloud in-process, or sidecars beside the JVM

swipe the figure sideways, or tap expand for full screen

1/5

the Spring Cloud way

Spring Cloud puts cross-cutting concerns in the application: Config Server for configuration, Eureka or Kubernetes for discovery, Spring Cloud Gateway at the edge, Resilience4j for circuit breaking, Micrometer for metrics.

cross-cutting concerns as librariesMesh course P5