Too much choice is as bad as too little choice
I've been on the operations side of technology most of my career. I've learned too much choice doesn't scale.
I keep playing a Twitter interaction from the OpenStack days back in my head. It was the wild west of “developer choice.” I made the argument that there was too much choice. Sure the public cloud enabled developers to deploy infrastructure to get to the business solution quickly. However, I argued that creating snowflakes makes it extremely difficult to manage applications centrally, what I call the new enterprise scale.
Fast forward to 2022, and app mobility is a genuine concern for many CTOs. Sometimes, GCP is a better platform for a given business process than AWS. However, you can’t simply move an application and its data from one cloud provider to the next. Kubernetes, you say? Remember, we’ve given app developers the green light to move as fast as possible. Sure, you can use EKS for your pods, but what about monitoring? How hard is it to move an app when operations are based on AWS Cloud Watch vs. Garfana?
Part of my holiday break research is looking deeply into application portability beyond Kubernetes Kubctl. What framework, if any, can an architect use to identify the dependency of a given application?