Customer Problem: If a user is upgrading their openshift cluster and there is a single Quay pod deployed and when the node is drained for upgrade, Quay pod will be rescheduled to run on a different node due to which there can be a downtime or a min or two
Even though HorizontalPodScaler exist it doesn't solve the problem under this scenario , so user has to edit the HPA to maintain min replicas to avoid downtime or scale up the deployment to have more pods so there is no downtime during openshift upgrade.
Both the solution are not acceptable to customer so we need a way to mention min replicas in quayregistry CR.
Goal: Avoid downtime when updating / reconfiguring Quay via the Operator or during rescheduling events
- when configured to be managed the HPA for Quay, Clair and Repo Mirroring Worker should be set to minimally 2