Managed means the control plane
A managed Kubernetes service runs the control plane for you: the API server, scheduler and etcd. You still supply and pay for the worker nodes, and you still own everything running on them.
That division explains most of the pricing differences. Some providers charge for the control plane, some give it away and make their margin on nodes.
Control plane pricing differs materially
- DigitalOcean and Linode include the control plane at no cost, so you pay only for worker nodes.
- Google Cloud charges a per-cluster management fee, with one zonal cluster free per billing account.
- AWS and Azure both charge per cluster per hour for the control plane.
For a small cluster this fee can be a large fraction of the total, which is why the developer clouds are attractive for modest workloads.
Maturity is not equal either
Google Kubernetes Engine is generally regarded as the most complete implementation, which is unsurprising given Kubernetes originated at Google. Autopilot mode goes further and manages the nodes as well.
AWS EKS is powerful and deeply integrated with the rest of AWS, at the cost of more setup. Azure AKS integrates with Entra ID and the Microsoft tooling. The developer clouds offer solid, simpler implementations that cover most needs without the surrounding complexity.
Ask whether you need it at all
Kubernetes solves problems that appear at a certain scale and organisational shape: many services, several teams, frequent independent deploys. Below that, it adds substantial operational overhead for benefits you are not collecting.
For a handful of services, a container platform or plain instances behind a load balancer is usually less work and cheaper to run. Adopting Kubernetes because it is the default answer is a real cost.
What to budget for beyond nodes
- Load balancers, which are provisioned per service of type LoadBalancer and billed hourly
- Persistent volumes, which outlive pods and keep billing
- Egress from the cluster, particularly across zones
- Log and metric retention, which grows with cluster size
Portability in practice
Kubernetes is genuinely portable at the workload level: a deployment manifest moves between providers with little change. What does not move cleanly is everything around it, including ingress controllers, storage classes, identity integration and load balancer annotations.
Plan for a migration to be a week of work rather than an afternoon, and it is a reasonable expectation.
Related accounts
Accounts on the platforms discussed here: Buy Google Cloud Accounts, Buy DigitalOcean Accounts, Buy Linode Accounts.
Related cloud accounts
Accounts on the platforms discussed here, with every configuration and price published up front:
- Buy Google Cloud Accounts — from $30.00. Vertex AI, Gemini models and BigQuery.
- Buy DigitalOcean Accounts — from $25.00. Droplets, managed Kubernetes and Spaces.
- Buy Linode Accounts — from $25.00. Akamai-backed compute and SMTP-capable accounts.
Further reading
- DigitalOcean vs OVHcloud: Which Cloud Account Should You Buy?
- Azure vs DigitalOcean: Which Cloud Account Should You Buy?
- Which Cloud to Use for AI and Machine Learning Work
- AWS vs Google Cloud: Which Cloud Account Should You Buy?
Official documentation
Provider behaviour, quotas and pricing are defined by the platform itself. These are the first-party references for the services discussed above:
Frequently asked questions
Usually DigitalOcean or Linode, because neither charges for the control plane and both have straightforward node pricing.
No. The control plane is managed; worker nodes, workloads, networking policy and upgrades of your own components remain yours.
Generally not. A managed container or app platform is less work until you have many services and several teams deploying independently.
