Historically, updates often result in some downtime of the component that is updated. In a system landscape with an increasing number of autonomous microservices that are updated independently of each other, recurring downtimes due to frequent updates of the microservices is not acceptable. Being able to deploy an update without downtime becomes crucial.
In this section, we will see how we can perform a rolling upgrade, updating a microservice to a new version of its Docker image without requiring any downtime. Performing a rolling upgrade means that Kubernetes first starts the new version of the microservice in a new pod, and when it reports as being healthy, Kubernetes will terminate the old one. This ensures that there is always a pod up and running, ready to serve incoming requests during the upgrade. A prerequisite for a rolling upgrade to work...
