Add per job cleanup methods to HighAvailabilityServices

The HighAvailabilityServices are used to manage services and persistent state at a single point. This also entails the cleanup of data used for HA. So far the HighAvailabilityServices can only clean up the data for all stored jobs. In order to support cluster sessions, we have to extend this functionality to selectively delete data for single jobs. This is necessary to keep data for failed jobs and delete data for successfully executed jobs.