DoD:

Currently we return early if we fail to apply a resource during installation https://github.com/openshift/hypershift/blob/main/cmd/install/install.go#L248

There's no reason why we wouldn't keep going, aggregate errors and return at the end.

It might help for scenarios where one broken CR prevent everything else from being installed, e.g.

https://redhat-internal.slack.com/archives/C02LM9FABFW/p1680599409023509?thread_ts=1680589848.540709&cid=C02LM9FABFW