+++ This bug was initially created as a clone of Bug #1796444 +++ Following the discussion here https://github.com/openshift/machine-config-operator/pull/1031#discussion_r371965677 we need to use the /readyz endpoint instead of the /healthz which we're currently using in gcp-routes-controller. Not fully aware of how reproducible this is and/or how to verify it (maybe just looking at the CI and our own tests)
This is related again to https://bugzilla.redhat.com/show_bug.cgi?id=1802534 and it's unclear how back we're going to backport, closing for now.