Zero-downtime deployments
In a mission-critical environment, the application must be always up and running. These days, we cannot afford downtime anymore. Kubernetes gives us various means of achieving this. Performing an update on an application in the cluster that causes no downtime is called a zero-downtime deployment. In this section, we will present two ways of achieving this. These are as follows:
- Rolling updates
- Blue-green deployments
Let’s start by discussing rolling updates.
Rolling updates
In the previous chapter, we learned that the Kubernetes Deployment object distinguishes itself from the ReplicaSet object in that it adds rolling updates and rollbacks on top of the latter’s functionality. Let’s use our web component to demonstrate this. We will have to modify the manifest or description of the deployment for the web component.
We will use the same deployment definition as in the previous section, with one important difference...
