When apiserver is not available briefly, the CEO status suggests that the openshift-etcd namespace does not exist:

> EtcdEndpointsDegraded: EtcdEndpointsController can't evaluate whether quorum is safe: CheckSafeToScaleCluster failed to get bootstrap scaling strategy: failed to get openshift-etcd namespace: namespace "openshift-etcd" not found...

which is evidently wrong, because the api has just responded with that exact status, which means etcd has returned this as a result.    

We should reword this message to indicate the actual error.

Code path is here:
https://github.com/openshift/cluster-etcd-operator/blob/release-4.17/pkg/operator/ceohelpers/bootstrap.go#L120-L123

any supported release 

always    

    1. created some temporary downtime on apiserver
    2. check the CEO status using oc get co    

CEO returns "failed to get openshift-etcd namespace: namespace "openshift-etcd" not found."

CEO should not indicate that namespace is missing