Application High Availability

Build your application components (for example, web framework, messaging tier, or datastore) to scale and always scale them to at least two pods. OpenShift Dedicated’s default scheduler configuration will ensure those two pods end up on separate compute nodes. This is the minimum requirement to ensure an upgrade or maintenance update can work on one node at a time and not cause any application component downtime. Health and liveliness checks must also be in place to prevent outages.

If an application component needs to satisfy a minimum throughput requirement that cannot be met with only one pod, you must scale that component to the point where removing one pod will leave enough compute power to satisfy demand. Using horizontal pod autoscaling is also a good way to ensure you always have enough capacity. Learn more about high availability (H/A) for applications.