Use case

Consider trigger based event correlation as described here: https://www.zabbix.com/documentation/4.0/manual/config/event_correlation/trigger

Note that this use case is practical for log and trapper items with Multiple problem event generation mode enabled. 

The goal is to reduce the trigger flapping in case the application stops and restarts happen in relatively short time causing no or very short service interruptions. 

Currently available trigger functions are evaluated immediately when e.g. problem line appears in the log and the problem is closed when the recovery expression is satisfied with all the tags and conditions. However, the practical requirement is to put the trigger in the actual Problem state only if in e.g. 5 seconds after the "problem" line is detected the recovery expression with matching tag value has not been found in the log. In this case a short disturbance in the service would not be triggered. 

Should that be implemented in a trigger function as a parameter or a configurable box in the frontend and API does not matter that much although a box in the web interface would be more user friendly and not overcomplicate already option-rich trigger functions.