The Modern Cloud: Where Your Time Goes to Die (and Your Sanity).

Kubernetes logo

Thousands of people, perhaps tens of thousands, work on essentially identical setups across the world. The ecosystem built around the added complexity of cloud-native technology has to sustain itself at this point, as people livelihoods depend on it. There are multi-million euro revenue companies (in Sweden only) that sell Managed Managed Kubernetes, which proves that even the managed solutions don't solve the problem for many.

Most of the complexity in Kubernetes comes from driving change after the initial setup. Each new service deployment, version upgrade and incident involves hour(s) of valuable engineering time, even for a skilled team. This is because Kubernetes is not a platform, rather it's made to built platforms on top of, and that's not how it's used in general.

In essence, companies pick technology that they believe will let them thrive in their respective markets. Those choices can be based on previous experiences, but realistically it's often based on the experiences of others. Considering there are monetary advantages in building interest in a more complex technology (Kubernetes in this case), I don't see much interest in tech companies driving change back towards simplicity.

I am still waiting for The One open-source project to save us from the shackles of Kubernetes and Cloud-Native technology. As someone who has build platforms on top of Kubernetes since ~2016, I don't believe the environment has improved much since then. What was hard then is still hard now, with some clear exceptions due to the excellent work from Cilium!

I'm hoping we can be that change at molnett. It all started from my wish of building an alternative to Kubernetes with simplicity and ease of operations in mind, and I'm still dead set on delivering on that one day.