In the process of replatforming legacy applications onto Google Kubernetes Engine, your organization aims to decompose a monolithic application into microservices without rewriting the application code. There are multiple instances with read and write access to a configuration file stored on a shared file system. How can you minimize the effort required to manage this transition?