I have a TEST Cluster. It's not similar to the Production one but I can do some kind of a/b testing between standard and logsdb index format and see the performance hit.
If you think it could be useful I can work on that tomorrow.
1 Like
Yes that could be informative to help understand the issue.... BUT if the HW Profiles are substantially different say like SSD vs the HDD on prod it may not as much help...
It will only be useful if there are settings / adjustment that could be made to help your Production cluster, which I am not sure what other setting on the Logsdb side would have a positive / material affect.
or on the merge side (be careful with these settings)
Lets see if / when the others weigh in before "we" add work to your plate. 
If you see the abnormal CPU usage, then hot thread dumps could be helpful to see what is the cause.