When Spring Cloud Config is in use, changes of configuration from GIT are read only during startup of microservice. It's also possible to manually call /actuator/refresh endpoint, when you use Actuator.
Example to propagate such changes in realtime involves Spring Bus Cloud, which requires installation of RabbitMQ, Kafka, which for me looks like overkill to make such a simple thing. It also requires to add queue dependency in every service.
When I use Eureka or other discovery service, all instances are registered and easily available. The missing bit is to call /actuator/refresh on instance, which has change in configuration.
Do I miss something, or should I write something like custom solution described here? https://naumoski.co.za/2017/12/20/spring-boot-config-server-and-client-with-refreshing-context/