Part 5 Managing applications at scale
The final part focuses on managing large numbers of pods that run diverse types of applications and workloads. In Kubernetes, you don’t create Pod objects directly. Instead, you create higher-level objects that are managed by controllers, which in turn create and maintain the underlying pods automatically. The controller you choose depends on the type of application you need to run.
In chapter 14, you’ll learn how a ReplicaSet manages a set of pod replicas. A ReplicaSet ensures that the desired number of application instances are always running. If a node fails, any pods belonging to the ReplicaSet that were running on that node are automatically recreated on other nodes, helping keep your application highly available.
Chapter 15 explains why you typically use a Deployment rather than creating a ReplicaSet directly. Kubernetes automatically generates the underlying ReplicaSets as needed, enabling the Deployment to provide powerful features such as automated, controlled rollouts and rollbacks of application updates.
Chapter 16 covers how to deploy stateful applications using the StatefulSet object. Unlike Deployments, which are designed primarily for stateless workloads, StatefulSets provide stable network identities, ordered scaling, and other behaviors specifically tailored to stateful applications.