Each instance we run is usually a live version of a specific image (Amazon AMI or Docker), which are themselves a snapshot of the code at a certain moment. We v

Each instance we run is usually a live version of a specific image (Amazon AMI or Docker), which are themselves a snapshot of the code at a certain moment. We version with Semver, so our machine-images and the corresponding servers all inherit those names. So for example: * You could find a github tag in our repo marking v1.1.1 * You could find a server image called the same (mydomain.com-v1.1.1) * If we're running that image as a live server, it too would be called mydomain.com-v1.1.1. * If we're running multiple servers off the same image in parallel, we'd just append an arbitrary UID to the end of the name to uniquely identify each instance. Hope that all made sense.