Currently when a trigger is deleted information about it is written into the "housekeeper" table.
Later housekeeper uses that data to remove data from "problems" table related to this trigger.
But "events" table is not cleaned up and events in it remain until the global housekeeping settings from Administration -> (General) -> Housekeeping -> Trigger data storage period
are reached. By default this is on year.
An installation with many changes (hosts removed etc) could end up with a huge amount of entries in events table that cannot be used in any way.
This is caused by ZBX-12975, if housekeeper deletes the events before configuration has been reloaded (trigger still exists in configuration cache), this can cause problems.

I believe this process could be improved by removing such orphaned events. For example 24h after the trigger has been deleted. 24h to be on the safe side that configuration has been reloaded.