A handful of failures on aws lately, very rare, but they do indicate that the logic we currently have in the test can trip when we don't really care.

https://prow.ci.openshift.org/view/gcs/origin-ci-test/logs/periodic-ci-openshift-release-master-ci-4.11-upgrade-from-stable-4.10-e2e-aws-upgrade/1534289681716350976

https://prow.ci.openshift.org/view/gcs/origin-ci-test/logs/periodic-ci-openshift-release-master-ci-4.11-upgrade-from-stable-4.10-e2e-aws-upgrade-workload/1533156263876104192

https://prow.ci.openshift.org/view/gcs/origin-ci-test/logs/periodic-ci-openshift-release-master-ci-4.11-upgrade-from-stable-4.10-e2e-aws-upgrade/1531959480453959680

We allow 10min * num revisions today, and most of the half dozen fails this month are within 5s - 1 minute. One was 25 minutes.

Extending to 15min per revision should handle these outliers and preserve our ability to detect massive changes.