Actuator, observability, and operations

Production readiness begins with user health and controlled management.

Actuator

Spring Boot Actuator provides management endpoints and infrastructure for health, metrics, auditing, and operations. Expose only what operators need. Liveness should mean the process is locally recoverable; readiness should mean the instance may receive traffic.

Observability

Micrometer provides vendor-neutral metrics and observation. Use OpenTelemetry or compatible tracing integration for distributed requests. Correlation IDs connect logs, traces, and messages.

Start with request rate, error rate, duration, and saturation. Add business signals such as completed orders and payment failures. Define SLIs and SLOs from user journeys, then alert on error-budget burn.

Logging

Use structured logs with event name, severity, service, environment, trace ID, and safe business identifiers. Never log credentials, bearer tokens, full payment data, or unnecessary personal information.

Graceful shutdown

Stop accepting new traffic, complete bounded in-flight work, pause consumers, and close pools. Ensure the deployment platform grants enough termination time.

Feynman check

Actuator is the service panel, metrics are gauges, traces follow one parcel, logs are the diary, and an SLO says how often the customer journey must work.