Smallest unit of compute you can define and manage.

• A. Node
• B. Pod
• C. Container
• D. ReplicaSet

Answer: (D)

The smallest unit you manage for scalable, reliable compute in Kubernetes is a ReplicaSet. It defines a template for identical pods and the exact number of replicas you want running. The ReplicaSet is responsible for creating and maintaining that number of pods, so if any pod dies or becomes unavailable, the ReplicaSet automatically replaces it to meet the desired count. That means it provides the basic control loop needed for replication and self-healing across the cluster.

A Pod is the actual execution unit that runs containers, but on its own a Pod doesn’t guarantee multiple instances or steady availability. A Node is simply a worker machine that runs pods, and a Container is the runtime inside a pod. In practice, deployments manage ReplicaSets to handle updates and scaling, but the core unit that encapsulates the idea of “defined compute with a desired replica count” is the ReplicaSet.