KatelloEventQueue involves the following operations

1) Pick an event thats not in-progress and mark the rest of the duplicate events below it as "in progress"
2) Do the operation
3) Delete all the duplicates

In a large scale environment where over 10's of 1000s of events (with many duplicates) gets queued up in the EventQueue we need a quicker way to handle all 3

The last part seems to delete 1 event at a time instead of deleting the duplicates en masse


To test this
1) create activation key with subs attached
2)Do simultaneou/multiple registers and unregisters in quick succession  in the 100s , 
3) sudo su - postgres; psql foreman; select count(*) from katello_events
make sure you have a lot of import_pools running
4) Notice that duplicate events are not immediately cleared after the first event pool. As in they happen slowly


other approach to repro this scenario

1) Stop foreman-tasks
2) foreman-rake console
3) Katello::Pool.first.id
>> 3 #note down the id
4) 30000.times {   ::Katello::EventQueue.push_event(::Katello::Events::ImportPool::EVENT_TYPE, 3) }

exit
5) sudo su - postgres; psql foreman; select count(*) from katello_events
make sure you have a lot of import_pools running
6) start foreman-tasks
7) Keep monitoring the counts in katello_events

It takes a long time to clear all the events and go to 0 (5 min in our demo sat)