As you can see from the attached monitoring, there is a different behaviour between saturday (restart of our core system including version 2.7.5_03 together with adding some more indexes, indexing on startup, you can see the escalation towards yesterday evening) and the startup yesterday. The peek at around 23:00 is because of an automated database backup, not critical.
between saturday and yesterday we had always CPU peeks on GC what we do not have since startup yesterday evening.
also the lower memory consumption is much higher from saturday compared to yesterday startup
may be the reindexing on startup has a problem?
we just added simple indexes:
- Action.created
- Action duration
- Action.descriptiontText
- Action.detail
- ObjectNode.created
- ObjectNode.internalReference
- ObjectNode.currentLocation
- ObjectNode.precessors