Certified Kubernetes Application Developer (CKAD) Study Guide, 2nd Edition

Chapter 10. Deployments

A big selling point of Kubernetes is its scalability and replication. To support those features, Kubernetes offers the Deployment primitive. In this chapter we’ll show the creation of a Deployment scaled to multiple replicas, how to roll out a revision of your application, how to roll back to a previous revision, and how to use auto-scalers to handle scaling concerns automatically based on the current workload.

Coverage of Curriculum Objectives

This chapter addresses the following curriculum objective:

• Understand Deployments and how to perform rolling updates

Working with Deployments

The primitive for running an application in a container is the Pod. Using a single instance of a Pod to operate an application has its flaws—it represents a single point of failure because all traffic targeting the application is funneled to this Pod. This behavior is specifically problematic when the load increases due to higher demand (e.g., during peak shopping season for an e-commerce application or when an increasing number of microservices communicate with a centralized microservice functionality, e.g. an authentication provider).

Another important aspect of running an application in a Pod is failure tolerance. The scheduler cluster component will not reschedule a Pod in the case of a node failure, which can lead to a system outage for end users. In this chapter, we’ll talk about the Kubernetes mechanics that support application scalability and failure tolerance. ...