You support a production service that runs on a single Compute Engine instance . You regularly need to recreate the service by deleting the crashing instance and creating a new instance based on the relevant image. You want to reduce the time spent on manual operations while following Site Reliability Engineering (SRE) principles . What should you do?