Currently logic for determining what to do with an analysis' results and sending events to monitoring are tightly coupled. Monitoring calls should instead be made explicit and decoupled from other logic