It also struck me that some projects such as Kubernetes (add to this many common programming languages and other tools) have equally become so complex, that we need to add dozens of other tools and dependencies to use them effectively. There were so many sponsors, exhibitors, and companies involved with KubeCon EU it took me around 15 seconds to scroll through the whole list.

Similarly, the infamous Cloud Native Computing Foundation (CNCF) landscape image has become so large you need to zoom out to see it all at once. There is a special interest group now tasked with making the landscape easier to navigate.

After a long era of (relatively) affordable cloud computing resources, there has been a temptation to keep throwing more cloud-based resources at problems. There was minimal impact on cost, and we gained a lot of advantage from the flexibility offered.

We didn’t want to write boilerplate code repeatedly, so we used frameworks and dependencies to (hopefully) help focus on “what was really important.”

We needed tools to handle demand, test and build applications, orchestrate infrastructure, run continuous processes, etc and etc and etc…

Because of all this abstraction, we lost direct insights into our technical stacks and what was happening inside applications. We added more and more tools to help us understand what we had lost sight of. This includes observability, but there are also concentrated efforts within the CNCF and other foundations to help bring clarity to the security risks introduced through nests of dependencies, such as software bills of materials.

Currently, for a handful of different reasons, the financial impact of this growth of cloud services is now starting to hit budgets, with a growth of tools to help estimate, monitor, and streamline costs. There’s even an entire sub-foundation of the Linux Foundation, the FinOps foundation, that aims to help increase skills and awareness of techniques and practices around efficient cloud usage.