Handling infrastructure at scale

One of the most demanding tasks when running large-scale multi-cluster Kubernetes in the cloud is dealing with the cloud infrastructure. In some respects, it is much better than being responsible for low-level compute, network, and storage infrastructure. However, you lose a lot of control, and troubleshooting issues is challenging.

Before diving into each category of infrastructure, let’s look at some general cloud-level considerations.

Cloud-level considerations

In the cloud, you organize your resources in entities such as AWS accounts, GCP projects, and Azure subscriptions. An organization may have multiple such groups, and each one has its own limits and quotas. For the sake of brevity, let’s call them accounts. Enterprise organizations’ infrastructure requirements will exceed the capacity of a single account. It’s critical to decide how to break down your infrastructure into different accounts. One good...