I've setup a 2016 RDS farm with 1 RD Connection Broker (RDCB), 40 RD Session Hosts (RDH), no Web Access and no Gateway. There are about 4000 connections (mostly reconnects) per hour. The connections are following this way: Client -> RDH -> RDCB -> RDH. Since I've got just one RDCB, this is no HA installation and the session database is local, not shared.
80% of the redirects are taking 0-1sec as they should, 10% are taking 2-10sec and 10% are taking more than 10 seconds. The time delta is always between the event ids 800 (redirection request received, TerminalServices-SessionBroker\Operational) and 801 (redirection request processed). I've not been able to find out what's taking the RDCB so long to redirect some sessions.
I've checked:
- hardware resources
- debug & analytic eventlogs
Has anyone experience with this phenomenon? Or else has anyone an idea how to further analyze it so that I'm able to find the real culprit?
Please feel free to ask for more information, I'll be more than happy to provide it.