Description of problem:

If you manually modify the notebook-controller-deployment to zero pods, rhods-operator doesn't change it back to the default (1 pod).

This doesn't happen with odh-notebook-controller-manager and rhods-dashboard, where the number of pods is restored by the operator

Note: With default configuration, if you kill the pod it is immediately restored automatically. So, I don't see it as a blocker bug for 1.16.

Steps to Reproduce

1. Install RHODS 1.16 livebuild
2. Scale down notebook-controller-deployment to 0
3. Wait a few minutes
4. Verify if notebook-controller-deployment still has 0 pods

Actual results:

notebook-controller-deployment stays with 0 pods

Expected results:

notebook-controller-deployment is scaled up to 1 pod

Reproducibility (Always/Intermittent/Only Once):

Always

Build Details:

RHODS 1.16 livebuild from yesterday

Workaround:

Manually scale up deployment

Additional info:

Upstream bug: https://github.com/opendatahub-io/odh-manifests/issues/663

Upstream PR: https://github.com/opendatahub-io/odh-manifests/pull/664