Why Spring Cloud @RefreshScope Prevents BeanPostProcessor from Processing Beans

When Spring-managed classes annotated with @Service or @Component also have the @RefreshScope annotation, the custom BeanPostProcessor fails to process them.

The reason is that BeanPostProcessor only runs when the bean definition is not synthetic (mbd == null || !mbd.isSynthetic()). The RefreshScope infrastructure marks the bean definition as synthetic, causing the condition to be true and skipping processing.

The relevant Spring source method is org.springframework.beans.factory.support.AbstractAutowireCapableBeanFactory#initializeBean(String, Object, RootBeanDefinition) , which checks the synthetic flag before invoking post‑processors.

RefreshScope sets this flag in org.springframework.cloud.context.scope.GenericScope#postProcessBeanDefinitionRegistry by calling root.setSynthetic(true) on the bean definition.

Consequently, any BeanPostProcessor that relies on custom annotations will not be invoked for beans annotated with @RefreshScope, leading to failures if the processor expects to handle those beans.

To work around this, avoid combining @RefreshScope with BeanPostProcessor logic that depends on custom annotations, or customize the processing to account for synthetic bean definitions.