spring-boot-observability

Observability in Spring Boot combines Actuator endpoints for health checks, Micrometer for custom metrics (Counter, Timer, Gauge), and OpenTelemetry for distributed tracing. Enable the actuator dependency, configure OTLP export, and add custom metrics to capture application behavior at scale across services.

Liveness probes restart unhealthy containers; readiness probes route traffic only to ready instances. Spring Boot Actuator health groups separate these concerns, letting Kubernetes make independent decisions about container lifecycle and traffic distribution for production stability.

Configure OTLP exporters in Spring Boot to send spans to a tracing backend, then use context propagation standards (W3C Trace Context) to correlate requests across services. OpenTelemetry span customization lets you capture domain-specific events and dependencies in distributed traces.

Yes. Actuator supports role-based endpoint exposure and path customization. Separate public health endpoints for probes from sensitive metrics endpoints, use low-cardinality metric tagging to prevent cardinality explosion, and configure HTTPS and authentication for production-grade security.

Track built-in JVM metrics (memory, GC, threads), HTTP request metrics (latency, throughput), and custom application metrics (Counter for events, Timer for operations, Gauge for state). Use low-cardinality tags (service name, environment) to keep metrics queryable and cost-effective.

Configure Spring Boot graceful shutdown to drain in-flight requests while exiting health check groups mark the service unavailable. Emit final metrics and traces before termination to ensure observability data captures the shutdown sequence without gaps in production monitoring.