Specifically for Charm QA, runs triggered by users manually should be prioritized over those created by the CharmUpdateMonitor. Test Observer needs to support setting a priority on reruns.