Measure the number of features you could ship per hour while maintaining a highly available service, not how much time you spent in deployment and monitoring.

With Kubernetes you can update your application with a container image holding a tag, killing old and outdated, while maintaining an artifact record of what you did, so you could rollback easily. Your infrastructure is not affected by the user modifications, and what's more, when Kubernetes receives a desired state configuration, it takes actions continuously to ensure it stays stable.

In Kubernetes, applications can be colocated on the same machines without impacting each other. Clusters with Kubernetes are easy as all machines are the same in the cluster, therefore users are not affected during scaling. High-level fo abstraction APIs allow you to have a nimble team supporting big development units. Declarative configuration makes a system less prone to costly man-made mistakes.

Kubernetes opens up opportunities for fast scaling of both resources and clusters, therefore fostering use of Cloud computing. It also facilitates shifting to Service-Oriented and Microservice Architectures and provides development teams with a list of instruments such as pods, abstraction APIs, namespaces etc. to achieve it.