4 Zero Downtime Operations

Kubernetes can automate many operations like restarting your code on application failure, or migrating your application in the case of hardware failure, all which work to make your deployment more reliable without you needing to monitor it 24/7. These automated operations are one of the key value propositions of Kubernetes, and an essential step to taking full advantage of everything it has to offer.

Kubernetes can also help you update your application without outages and glitches by booting the new version of the application and monitoring its status to ensure it’s ready to serve traffic, before removing the old version.

To help Kubernetes help keep your application running without downtime during normal operations and upgrades, you need to provide certain information about the state of your application with a process known as Health Checks. In the next section we’ll go through adding the various health checks to your application, and in the later section, how these can be used with Kubernetes’ in-built roll out strategies to update your application without glitches or downtime.

4.1   Automated Uptime with Health Checks

4.1.1   Liveness and Readiness Probes

4.1.2   Adding a Readiness Probe

4.1.3   Adding a Liveness Probe

4.1.4   Designing Good Health Checks

4.1.5   Probe Types

4.2   Updating Live Applications

4.2.1   Rolling Update Strategy

4.2.2   Replacement Strategy

4.2.3   Blue / Green Strategy

4.2.4   Choosing a Rollout Strategy

4.3   Summary