Skip to main content
Welcome to this lesson from OpenShift 3 for the Absolute Beginners. In this article we cover how to scale applications in OpenShift — manually, from the CLI, and automatically (autoscaling). A DeploymentConfig (OpenShift) manages a ReplicationController; a Kubernetes Deployment manages a ReplicaSet. Both a ReplicationController and a ReplicaSet are responsible for maintaining multiple replicas of a Pod. Scaling an application means changing the number of replicas in the resource’s spec.replicas field. In demos so far we used a single replica. You can change the replica count by editing the object’s YAML, but the OpenShift web console provides a quick interactive way to scale up or down with one click. Traffic is automatically distributed to replicas by a Service (for cluster-internal traffic) or by the Router (for external traffic).
A slide titled "Scale Deployment" showing a diagram of a deployment/replication controller with stacked container/pod icons on the left. On the right is a web console screenshot for a "simple-webapp" deployment showing details and a circular indicator of 5 pods.
You can scale interactively in the web console for quick changes. Use the oc CLI for repeatable, scriptable operations and automation.

​
Common methods to scale

This section shows common ways to scale a DeploymentConfig or a Deployment in OpenShift: using the web console, using oc to scale/edit/patch, and enabling Horizontal Pod Autoscaling.

​
1) Scale using the oc CLI

Scale a DeploymentConfig directly:
If you use Kubernetes Deployments instead of DeploymentConfigs:
After scaling, verify new pods:
Example output:

​
2) Edit the YAML directly

Open the resource in your editor, change the spec.replicas value, save and exit. OpenShift will reconcile and create or terminate pods as needed. Edit a DeploymentConfig:
Edit a Deployment:
Alternatively, apply a modified YAML file:
Ensure spec.replicas in the file contains the desired count.

​
3) Patch the resource (quick one-liner)

Use oc patch to update replicas without opening an editor:

​
4) Enable autoscaling (Horizontal Pod Autoscaler)

OpenShift supports Horizontal Pod Autoscaling for DeploymentConfigs and Deployments. The oc autoscale command creates a HorizontalPodAutoscaler that scales based on metrics such as CPU utilization. Example:
This instructs the autoscaler to keep average CPU utilization near 80%, scaling between 2 and 10 replicas as needed. Check existing autoscalers:

​
Commands quick reference

When editing live resources, be mindful of production traffic. Changing replicas may disrupt session-affine workloads unless you use sticky sessions or an external session store.

​
Notes on traffic distribution

  • A Service that fronts your pods will automatically load-balance traffic across available replicas.
  • For externally exposed applications, the Router (OpenShift’s ingress point) distributes requests across Service endpoints (the pods).
  • Ensure readiness probes are configured so traffic is only sent to healthy pods.

​
Troubleshooting tips

  • If new pods remain in Pending, check node capacity with oc get nodes and pod events with oc describe pod <pod-name>.
  • If pods start but never become Ready, inspect logs and probe configurations:
    • oc logs <pod-name>
    • oc describe pod <pod-name>
  • For HPA issues, confirm metrics are available (metrics-server or Prometheus adapter as required).

​
Summary

  • Scaling changes the spec.replicas value of a DeploymentConfig or Deployment.
  • Web console scaling is convenient for ad-hoc changes; the oc CLI is best for automation.
  • Use oc scale, oc edit, or oc patch to adjust replicas from the CLI.
  • Use oc autoscale to add Horizontal Pod Autoscaling based on CPU (or other supported metrics).
  • Services and the Router handle traffic distribution across replicas automatically.

Watch Video