At first glance, a pod seems pretty similar to a container, but the main difference is that we can have multiple containers running inside a single pod. We can think of a pod as a way to group containers that cooperate to do something.

In most cases though, especially for simple applications like the ones we are running here, we will have only one container per pod.

When we run more than one container in a single pod, it’s usually to support the primary application. For example, we could have our primary nginx container running alongside another container that periodically pulls a github repository to update the website that nginx is serving.

Multi-container pods

In this case, the primary container is nginx, which is serving our website, and the container that is getting the code from github is there just to help this pod serve its main function. 

Does that mean that multi-container Pods are useless? They’re not. There are scenarios when having multiple containers in a Pod is a good idea. However, they are very specific and, in most cases, are based on one container that acts as the main service and the rest serving as side-cars.

A frequent use case is multi-container Pods used for:

Continuous integration (CI)

Continuous Delivery (CD)

Continuous Deployment processes (CDP)

We’ll explore them later. For now, we’ll focus on single-container Pods.