Currently the PassiveStatusMonitor triggers poller status changes (up/down) at the time the monitor is triggered. While this makes sense for most monitors, given that this feature is driven by events sent into the system, this is innaccurate in cases where the event is delayed in being received by the core. If the trap/syslog/event processing is able to handle extracting the original timestamp of the message to record on the event, allowing the PassiveStatusMonitor to "backdate" the serviceDown event would provide more accuracy to outages and in turn give customers a more accurate SLA calculation.

Example: An event is firing from an external system and it is sending using uei.opennms.org/services/passiveServiceStatus as uei with all the relevant params. <time> is also sent in the event xml. What can be seen is that the passiveServiceStatus event gets the correct timestamp, but when pollerd raises a new event, it uses "current" time when it raises it. This is tricky if we have a situation where we get those initial events "late".

For example, if some 3rd party system sends an event with timestamp 2022-04-05 13:34:42+0000, but it actually arrives "now". It would be more accurate to use original timestamp as eventtime, and have Pollerd's eventtime set same as the original timestamp. This way we could get correct timestamps to SLA report.