Remember when serverless was supposed to make Kubernetes irrelevant? Write a function, let the cloud handle everything, no more YAML hell. That was the pitch five-plus years ago. It didn't happen — and the way it didn't happen is more interesting than a simple "containers won."
What actually happened: AWS Fargate, Cloud Run, and similar platforms quietly merged the two models. You write and deploy like serverless, but underneath it's containers doing the work. The industry didn't pick a winner — it absorbed serverless's developer experience into container infrastructure and called it a day.
Here's the part that should bother people more than it does: this means the "choose serverless for simplicity" argument from 2020 was really an argument for better tooling, not a different execution model. Teams that migrated entire architectures to FaaS chasing operational simplicity could often have gotten the same simplicity with a managed container platform and more control to show for it.
My actual take: the serverless vs. containers debate was never really about technology. It was about who manages the complexity — the platform team or the cloud provider. Platform engineering teams becoming standard practice is the real signal here: orgs are choosing to own that complexity again, just with better abstractions than bare Kubernetes used to offer.
Genuinely curious — did your team ever fully commit to serverless-first, and if so, did it actually simplify things or just move the complexity somewhere else?
#AWS #Serverless #Kubernetes #CloudComputing #DevOps #PlatformEngineering #CloudArchitecture #CloudNative #TechDebate #BuildOnAWS
Enjoyed reading this content? Let the author know!
Your likes, comments, shares, and saves help creators reach more builders.
Loading recommendations
Loading article