These basic Objects may then be built upon to form higher-level Objects:

• ReplicaSet: Manages a set of Pods so that a specified number of replicas are maintained within the cluster.
• Deployment: An even higher-level abstraction than ReplicaSet, a Deployment Object will manage a ReplicaSet to ensure that the right number of replicas are running, but also allows you update your configuration to update/deploy a new ReplicaSet.

• StatefulSet: Similar to a Deployment, but in a Deployment, when a Pod restarts (for example, due to scheduling), the old Pod is destroyed and a new Pod is created. Although these Pods are created using the same specification, they are different Pods because data from the previous Pod is not persisted. In a StatefulSet, the old Pod can persist its state across restarts.
• DaemonSet: Similar to ReplicaSet, but instead of specifying the number of replicas to run, a DaemonSet is intended to run on every node in the cluster.
• Job: Instead of keeping Pods running indefinitely, a Job object spawns new Pods to carry out tasks with a finite timeline, and ensures that the Pods terminates successfully after the task completes.

The aforementioned higher-level Objects rely on the four basic Objects.