Skip to content

CONCEPT Cited by 1 source

Managed Kubernetes service

A managed Kubernetes service is a hosted offering where the cloud provider operates the Kubernetes control plane — and, in stronger variants, the data plane too — on behalf of the customer. The customer interacts with the Kubernetes API and deploys workloads using standard K8s object types (Deployments, Services, etc.), but does not own the etcd cluster, the API server, the scheduler, Kubelet patching, or (in nodeless variants) Node capacity planning.

Spectrum

Tier Control plane Data plane Example
Self-managed K8s operator operator kubeadm, kops
Managed control plane only provider operator vanilla EKS, vanilla AKS, vanilla GKE
Managed data plane provider provider (with operator-owned Nodes) EKS Auto Mode, GKE Autopilot
Nodeless provider provider (no Node object) Fly Kubernetes (FKS), Fargate-on-EKS

Fly.io's "blessed path" framing

Fly.io's FKS beta post positions FKS as their "blessed path"™️ to using Kubernetes backed by Fly.io infrastructure" — a deliberately opinionated managed service:

We take care of the complexity of operating the Kubernetes control plane, leaving you with the unfettered joy of deploying your Kubernetes workloads.

Fly's nodeless composition pushes beyond EKS / GKE / AKS defaults by also eliminating the Node / capacity-planning surface, via the Virtual-Kubelet + Fly Machines substrate.

Operator responsibilities by tier (typical)

  • Self-managed: etcd HA, control-plane upgrades, Kubelet patching, Node autoscaling, CNI operations, CSI driver choice, cluster DNS operations.
  • Managed control plane: Node autoscaling (Karpenter / cluster autoscaler), Kubelet AMI upgrades, CNI/CSI choice.
  • Managed data plane: maintenance-window consent, Pod Disruption Budgets, workload compatibility with managed-AMI refreshes.
  • Nodeless: workload-level concerns only (Pods, Services, HPA…); provider owns everything below the Pod.

Seen in

Last updated · 200 distilled / 1,178 read