In Standby mode, a particular instance of a polling notification will execute the notification according to its configured schedule. When you start the cluster, the polling notification that executes the first scheduled run is considered to be the primary notification. This instance will continue to execute the scheduled run as long as it is enabled and fully functional. If at any time this notification becomes disabled, another notification in the cluster will assume control. The notification that assumes control is arbitrary. After a notification has control, it will continue to execute the schedule for as long as it is enabled and fully functional.

In Distributed mode, any instance of the polling notification can execute the currently scheduled run. The notification that executes the current scheduled run is arbitrary. If a notification does not complete executing within the amount of time specified in the Max Process Time field, the system considers that notification to be "dead." (For details about Max Process Time, see Notification-Specific Settings in Settings.) Another (enabled) instance in the cluster will recognize this situation and will attempt to execute the scheduled run.