By default, scale-out operations performed manually or by cluster autoscale rules require the allocation and provisioning of new nodes, and scale-in operations delete nodes. Scale-down mode is a relatively newer concept that allows us to choose whether to delete or deallocate nodes. Having the ability to deallocate, rather than delete, nodes is a major performance benefit, as the time it takes to spin up new nodes will be significantly decreased. You will not be charged when nodes are deallocated.

In this blog post, we will attempt to explain the current storage options that exist in Kubernetes. If you are new to Kubernetes, learning about its capabilities of managing the application state can be a daunting task. Container images are built-in layers, with the runtime layer being writable. However, any files on this writable layer are only available for the container’s lifetime. We can mount a volume to a directory inside the container to have persistent data.