AutoscaleSustaining tests fail during upgrades. This failure is intermittent. The goal is to investigate if we have insufficient resources for running the tests which causes the autoscaler to not scale the application enough, or whether upgrades from Serving 0.19 to 0.20 on OpenShift introduce an issue which is caught by the test.

Example: https://prow.ci.openshift.org/view/gs/origin-ci-test/pr-logs/pull/openshift-knative_serverless-operator/803/pull-ci-openshift-knative-serverless-operator-master-4.6-upgrade-tests-aws-ocp-46/1360189729135398912