No matter how much we plan, we will never manage to have a cluster with capacity that should serve us equally well today as tomorrow. Things change, and we'll need to be able to adapt to those changes. Ideally, our cluster should increase and decrease its capacity automatically by evaluating metrics and firing alerts that would interact with kops or directly with AWS. However, that is an advanced topic that we won't be able to cover. For now, we'll limit the scope to manual cluster updates.
With kops, we cannot update the cluster directly. Instead, we edit the desired state of the cluster stored, in our case, in the S3 bucket. Once the state is changed, kops will make the necessary changes to comply with the new desire. ...
