Learn OpenShift

Autoscaling your application depending on CPU and RAM utilization

You can scale pods in your application using the oc scale command, but it has two disadvantages:

• It has to be run manually every time you need to scale a pod up or down
• You have to take into account CPU and RAM utilization yourself

This approach doesn't allow businesses to adapt quickly to constantly changing customers demands. There is a better way—HorizontalPodAutoscaler.

At the time of writing, autoscaling can only track CPU and RAM usage. Traffic-based autoscaling, for instance, isn't supported.

Let's login as system:admin and see if Hawkular, Cassandra, and Heapster pods are up and running:

# oc login -u system:admin...<output omitted>...# oc get po -n openshift-infra ...