Description of problem:

There is a case where the hypershift operator installation fails half way through where the operator deployment is installed but the rest of the hypershift operator installation fails. The agent's HO installer is designed to re-try every 2 minutes but it also checks if the deployment is already installed with the correct set of HO images. In this case of half successful installation, the re-try does not work because the HO deployment is already done.

Also in this case, the addon does not add the "hypershift.open-cluster-management.io/createBy" annotation to the hypershift operator so deleting the addon does not remove the hypershift operator.

Version-Release number of selected component (if applicable):

How reproducible:

Steps to Reproduce:

1. Follow https://github.com/stolostron/hypershift-addon-operator/blob/main/docs/advanced/enable_hypershift_operator_RHOBS.md to install the hypershift operator via the hypershift addon but skip the pre-req step to make the installation fail.
2. Look at the hypershift addon log. 

2023-02-22T15:03:24.441Z	INFO	agent.agent-reconciler	install/upgrade.go:70	check if HyperShift operator re-installation is required (startup=true, installfailed=false)
2023-02-22T15:03:24.441Z	INFO	agent.agent-reconciler	install/upgrade.go:75	hypershift operator re-installation is required
2023-02-22T15:03:24.445Z	INFO	agent.controller-manager-setup	agent/agent.go:199	the maximum hosted cluster count set to 80
2023-02-22T15:03:24.445Z	INFO	agent.controller-manager-setup	agent/agent.go:200	the threshold hosted cluster count set to 60
2023-02-22T15:03:24.446Z	INFO	agent.agent-reconciler	install/hypershift.go:172	hypershift operator exists but not deployed by addon, skip update 

Actual results:

The hypershift addon does not try to install the HO again.

Expected results:

The hypershift addon should try to install the HO every 2 minutes until it is successful.

Additional info: