Kubernetes Patterns, 2nd Edition

Part II. Behavioral Patterns

The patterns in this category are focused on the communications and interactions between the Pods and the managing platform. Depending on the type of managing controller used, a Pod may run until completion or be scheduled to run periodically. It can run as a daemon or ensure uniqueness guarantees to its replicas. There are different ways to run a Pod on Kubernetes, and picking the right Pod-management primitives requires understanding their behavior. In the following chapters, we explore the patterns:

• Chapter 7, “Batch Job”, describes how to isolate an atomic unit of work and run it until completion.

• Chapter 8, “Periodic Job”, allows the execution of a unit of work to be triggered by a temporal event.

• Chapter 9, “Daemon Service”, allows you to run infrastructure-focused Pods on specific nodes, before application Pods are placed.

• Chapter 10, “Singleton Service”, ensures that only one instance of a service is active at a time and still remains highly available.

• Chapter 11, “Stateless Service”, describes the building blocks used for managing identical application instances.

• Chapter 12, “Stateful Service”, is all about how to create and manage distributed stateful applications with Kubernetes.

• Chapter 13, “Service Discovery”, explains how client services can discover and consume the instances of providing services.

• Chapter 14, “Self Awareness”, describes mechanisms for introspection and metadata injection into applications.