AOP, events, scheduling, and caching

Spring adds cross-cutting behavior around beans through proxies and supports events, tasks, schedules, and cache abstractions.

AOP

Aspect-oriented programming is useful for concerns applied consistently across many methods: transactions, security, metrics, tracing, or narrowly defined logging. A pointcut selects join points; advice runs before, after, or around them.

Avoid hiding core business flow in aspects. If reading the method cannot reveal that an order is published or a fee is charged, the design is surprising.

Events

Application events decouple in-process components. By default, listeners run in the publisher's thread. @TransactionalEventListener can bind handling to a transaction phase. In-process events disappear if the process crashes; use a durable broker/outbox for cross-service guarantees.

Scheduling and async

@Scheduled triggers recurring work. In multiple replicas, every replica may run it unless coordination exists. @Async uses an executor; configure pool size, queue, rejection, context, and exception handling.

Caching

@Cacheable can avoid repeated work, but every cache needs a key, TTL, invalidation, maximum size, consistency expectation, and failure behavior.

Feynman check

A proxy is a gate adding checks. An event is a bell inside one building. A message broker is a postal service between buildings. A cache is a whiteboard copy that can become stale.