So, I was looing for information on how much of an impact EV 10 may have on Exchange Server 2010 disk queue length and came accross the forum discussion below.
https://www-secure.symantec.com/connect/forums/ev9...
Interestingly, the 32-bit MonitoringAgent was mentioned and that in EV 9 it suffers from a significant memory/handle leak. The user who started the discussion mentioned the process was using 224 MB and 15K handles. I decided to check our EV server - 883 MB and 60,830 handles.
I killed the process and when it respawned, after 10 minutes it is only using 22 MB and 795 handles. I will be watching it over the next few days but it seems like this is still an issue, and a significant one at that.
In case it is related, I will mention that we use Symantec BackupExec 2012 R2 to backup our EV server. BackupExec/EV has consistently had a problem taking the indexes and vault stores out of backup mode when the backup completes, mainly after the weekend full. We've been dealing with this issue literally for the last four years. I had to manually remove one of the vault stores and indexes from backup mode yesterday morning and of course there was a backlog of items not yet processed in the main journal mailbox (around 200 k).