Adding and removing an apiservice can break other components that rely on discovery (because they can't access a particular endpoint).

The failure here is oc new-app which exits because the discovery client failed, but in this case should be making forward progress because we don't need discovery info for all endpoints, just those that are in a new-app call (we would need to error if we ended up needing that info).

Failed here https://openshift-gce-devel.appspot.com/build/origin-ci-test/pr-logs/pull/18816/pull-ci-origin-e2e-gcp/2909/#conformanceareanetworkingfeaturerouter-the-haproxy-router-should-serve-the-correct-routes-when-running-with-the-haproxy-config-manager-suiteopenshiftconformanceparallel because the API aggregation test was running at the same time (registering wardle).

/tmp/openshift/build-rpms/rpm/BUILD/origin-3.11.0/_output/local/go/src/github.com/openshift/origin/test/extended/router/config_manager.go:43
Expected error:
    <*util.ExitError | 0xc422547a70>: {
        Cmd: "oc new-app --config=/tmp/admin.kubeconfig --namespace=e2e-test-router-config-manager-77v9b -f /tmp/fixture-testdata-dir572013924/test/extended/testdata/router-config-manager.yaml -p IMAGE=registry.svc.ci.openshift.org/ci-op-y5k4hf9x/stable:haproxy-router",
        StdErr: "--> Deploying template \"e2e-test-router-config-manager-77v9b/\" for \"/tmp/fixture-testdata-dir572013924/test/extended/testdata/router-config-manager.yaml\" to project e2e-test-router-config-manager-77v9b\n\n     * With parameters:\n        * IMAGE=registry.svc.ci.openshift.org/ci-op-y5k4hf9x/stable:haproxy-router\n\nerror: Unable to to get list of available resources: unable to retrieve the complete list of server APIs: wardle.k8s.io/v1alpha1: the server could not find the requested resource",
        ExitError: {


The aggregation test will be disabled from openshift e2es until this bug is fixed.

The metrics server has also been a source of this failure, the installer e2es were blocked for over a week due to flakiness in metrics server.