Kubernetes is genuinely excellent infrastructure — at the right scale. We've helped companies migrate to K8s, and we've also talked companies out of it.
The Case Against K8s for Small Teams
Kubernetes has significant operational overhead. Someone on your team needs to understand cluster management, node pools, ingress controllers, certificate management, persistent volumes, and resource quotas. For a team of 3-5 engineers, this overhead is often not worth the benefits.
When K8s Makes Sense
The threshold we use: you're running more than 8-10 containerised services that need independent deployment, scaling, and resource isolation. Below that, a managed container service will give you 90% of the benefit with 20% of the complexity.
The Middle Path
For teams that are K8s-curious but not K8s-ready, we often recommend managed Kubernetes with a tool like Helm for application packaging.
Real Thresholds
Under 5 services: Docker Compose is often fine. 5 to 10 services: use a managed container service. 10 to 20 services: use managed Kubernetes. 20 or more services: full Kubernetes with platform engineering investment.
The Cost Question
K8s clusters have a minimum viable cost — budget 25,000 to 40,000 per month minimum before adding actual application workloads.