Managed Kubernetes (OpenShift) · Container-only

Azure Red Hat OpenShift

Fully managed OpenShift clusters for running containers in production on Kubernetes.

When to use it

Choose Azure Red Hat OpenShift when your team already invests in OpenShift — its developer workflows, operators, and tooling — and wants a fully managed, jointly supported (Microsoft and Red Hat) cluster on Azure. It suits organizations standardizing on OpenShift across hybrid and multicloud.

When not to use it

Avoid Azure Red Hat OpenShift if you have no OpenShift requirement — plain AKS is a lighter, lower-cost managed Kubernetes. If you want containers without managing any cluster, Container Apps removes the orchestration overhead entirely.

Trade-offs at a glance

Hosting model Managed Kubernetes (OpenShift)
Container support Container-only
Minimum nodes 6 (3 primary + 3 worker)
State management Stateless or stateful
Web hosting Agnostic
Autoscaling Pod autoscaling and cluster autoscaling
Load balancer Azure Load Balancer or Application Gateway
Scale limit 250 nodes per cluster (default)
Multiregion Single region only — external router plus multiple clusters
Virtual network integration Supported
Hybrid connectivity Supported
GPU support Supported
TLS Supported
Architecture styles Microservices, event-driven architecture
Required skills OpenShift or Kubernetes administration
Operational overhead High — cluster management
Best for teams Teams with an existing OpenShift investment

Frequently asked questions

Who manages and supports Azure Red Hat OpenShift?

It's jointly engineered and operated by Microsoft and Red Hat, with integrated support, so the cluster, control plane, and updates are managed for you.

How is ARO different from AKS?

Both are managed Kubernetes, but ARO is OpenShift — it adds Red Hat's developer experience, operators, and opinionated platform on top, and is the right choice when you've standardized on OpenShift.

What is the minimum cluster size for ARO?

A cluster starts with at least three control-plane nodes and three worker nodes, so six nodes is the practical minimum for production.

Starting point, not a verdict

This is one candidate from the Azure compute decision guide. The right choice depends on your full requirements — evaluate scaling, cost, and operational fit before committing.

Browse all tools →