We are occasionally having client freeze ups in online mode and delays when sending (it sits there for several seconds). There RPC dialog notifications occasionally on the client as well. We have 10 database servers (some are DR DAG servers), 5 CAS/HT servers and 1 dedicated HT server. We have 4 servers in a hardware LB CAS array. We also have 2 DC's in the same site. These are all VM's.
We are moving to a shared desktop (citrix) non-VDI (not by choice) environment that requires online mode for the clients. Previously in cached mode there were no noticeable issues.
I've run the exchange troubleshooter repeatedly and it doesn't really appear there are too many highly active users. 6 users at any given time can account for 30% of the MAPI CPU utilization but everything appears to be healthy. Occasionally our DC's have high CPU utilization but that seemed to be due to LDAP lookups which I offloaded to another dedicated DC in another site. We do have BES but I have that going to a specific CAS server in our DR site.
I'm monitoring the servers in Solarwinds APM and not seeing anything that's out of spec other than our DAG's being behind in log files during the day.