We've often observed mod-source-record-manager module thrashing running OOM. On deeper investigation of what's causing this irrespective of giving plenty of resources, we observed that it's the sorting that causes the problem.
Steps to reproduce:
1. In Data Import app , load all log records (I tested with ~400 records)
2. Sort on "Ended Running" column multiple times - Sorting as much as 4-5 times is sufficient to get the module to crash irrespective of the amount of memory/CPU given to the module
Backend query sent :
Attached is the screenshot that shows the spikes in CPU and module crashes running OOM.
Here are the resources that are currently given to the module:
CPU - 5120
maxRAMPercentage - 85%
metaspace - 5632M
Xmx - 6144M
memory - 6800M
memoryReservation - 6144M