As Chris said this is probably related to the VPN more than the basic session timeouts, but try enabling TCP state logging to obtain more details about how and why your telnet connections are being terminated:
sk101221: TCP state logging
Is there some particular reason you are not synchronizing the telnet sessions across the cluster members? If deployed as shown in your screenshot a failover will cause all of the currently open telnet sessions to be killed on the newly active member.
Also note that the TCP connection timeout can now be increased well beyond 86400 seconds if you have the latest Jumbo HFA: sk168872: Virtual session timeout for a TCP Service (86400 seconds) is not long enough for a specifi...
Gateway Performance Optimization R81.20 Course
now available at maxpowerfirewalls.com