Oracle and Red Hat continue to expand Red Hat OpenShift on Oracle Cloud Infrastructure (OCI), helping you run enterprise Kubernetes workloads with more automation, flexibility, and operational consistency.

Today, Oracle and Red Hat are introducing autoscaling support for OpenShift on OCI, generally available for OpenShift 4.22 and later. Your clusters can now adjust worker capacity automatically as application demand changes, so you can run dynamic workloads without manually resizing worker pools.

You can enable elastic worker capacity when you create a new OpenShift cluster on OCI, or add it to a cluster you already run.

Why autoscaling matters

Application demand is rarely static. Development environments, batch processing, data services, AI workloads, and seasonal business applications all shift compute demand throughout the day.

Without autoscaling, you either size worker capacity for peak demand or add nodes by hand when workloads cannot be scheduled. Both increase operational effort, and the first leaves compute resources idle during quieter periods.

Autoscaling on OCI expands worker capacity when more compute is needed and contracts it when demand drops, within the limits you define.

How this works with OpenShift autoscaling

Autoscaling for OpenShift on OCI drives the standard OpenShift autoscaling machinery rather than replacing it, and operates alongside the autoscaling controls you already use.

When pods cannot be scheduled, the cluster autoscaler increases the replica count on the relevant machine deployment. Cluster API creates the corresponding machine and machine set resources, and the Cluster API Provider for OCI provisions the OCI compute instances behind them. You set the boundaries once, on an OCIClusterAutoscaler custom resource that carries minNodes and maxNodes, and the existing controllers do the rest.

Nothing about your current autoscaling configuration has to change. If you already run the OpenShift cluster autoscaler or machine autoscalers, they continue to work as they do today.

What you can do

Autoscaling for OpenShift on OCI fits into the deployment experience you already use. You can:

  • Enable autoscaling when you create a cluster.
  • Add autoscaling to a cluster you already run.
  • Set minimum and maximum worker capacity.
  • Choose OCI virtual machine or bare-metal shapes for autoscaling workers.
  • Use OCI Object Storage as the source for the worker image.
  • Monitor scaling activity with the OpenShift resources you already use, including the OCIClusterAutoscaler status, machine deployments, machine sets and machines, node readiness, and controller logs.

Autoscaling is integrated with the OpenShift on OCI Terraform stack, so you configure it alongside the existing cluster deployment and post-installation flows.

Enabling autoscaling

You can enable autoscaling in two ways.

  1. At install time. Turn on autoscaling while you create a new OpenShift cluster on OCI. Use this when elastic worker capacity should be part of the cluster from the start.
  2. On an existing cluster. Add autoscaling to an OpenShift cluster you already run on OCI, without rebuilding it.

Both paths use the OpenShift on OCI Terraform stack to collect the required OCI and OpenShift configuration.

Preparing the worker image

Autoscaling workers use a Red Hat Enterprise Linux CoreOS (RHCOS) image that matches your OpenShift cluster version. Download the matching RHCOS image, upload it to OCI Object Storage, and give the Terraform stack a read-only pre-authenticated request (PAR) URL.

For virtual machine workers, use the downloaded QCOW2 image directly.

For bare-metal workers, prepare the image with iSCSI boot configuration before you upload it. Run the helper script in the OpenShift on OCI quickstart repository from a Linux host or virtual machine.

Built for OpenShift operations on OCI

Autoscaling builds on the operational model you already use for OpenShift clusters. You continue to use OpenShift tooling to inspect cluster state, workloads, nodes, and events, while OCI provides the underlying compute, networking, and image services.

You keep a familiar operating experience and hand off the manual capacity management that workloads with changing demand would otherwise require.

Availability and support

Autoscaling for OpenShift on OCI is generally available and supported on OpenShift 4.22 and later.

The autoscaling operator is fully supported by Oracle. Open support requests through Oracle Support. Red Hat support for your OpenShift subscription is unaffected and continues to apply to the cluster.

Getting started

  1. Choose the OpenShift cluster that needs autoscaling, or create a new one running OpenShift 4.22 or later.
  2. Prepare the matching RHCOS worker image and upload it to OCI Object Storage.
  3. Use the OpenShift on OCI Terraform stack to enable autoscaling, either at install time or on an existing cluster.
  4. Configure the worker shape, image source, and minimum and maximum capacity.
  5. Apply the stack and verify that the autoscaling resources are available in OpenShift.

Detailed installation, validation, cleanup, and troubleshooting steps are in the project documentation.

Learn more