Your organization is migrating legacy applications to Google Kubernetes Engine and needs to decompose a monolithic application into microservices. Multiple instances currently read and write to a shared configuration file on a shared file system. To minimize transition effort and avoid rewriting application code, what should you do?