IcingadB 1.4.0 cleanup causes high database load #1003
Description
Describe the bug
A clear and concise description of what the bug is.
I've set up IcingaDB with a few thousand hosts and a postgres cluster as backend.
since the upgrade from 1.1.1 to 1.4.0, including applying all outstanding database migrations, icingadb runs regular cleanups of history tables, notification_history cleanup takes a very long time due to missing indexes on user_notification_history.
If I create the index manually (CREATE INDEX idx_user_notification_history_notification_history_id ON user_notification_history (notification_history_id);), the performance is back to normal and cleanups take just a few milliseconds. Is this index missing in migrations?
- Icinga DB version: 1.4.0
- Icinga 2 version: 2.15.0
- Operating System and version: CentOS Stream 9