We have an Always on VPN RRAS server (Server 2019 Std), which has been in place for 2yrs now without any issues, but over the last couple of weeks, we have started to experience mass client disconnects.
At any one time there will be 500-800 people connected, and suddenly 50-150 people will disconnect all at exactly the same time, then reconnect automatically a few seconds later.
Normally this would not be much of an issue, but the software the clients are using is very sensitive to network drops and it makes the issue more obvious.
When this happens, Event ID 20275 is logged on the VPN server - The user with IP address x.x.x.x has disconnected, but no other errors or events are logged that correspond with this time of disconnects.
All other users (there is a mix of device and users connections) remain unaffected, and so far no discernable pattern has been found for the users that do disconnect.
All users are running Windows 10 1909 or later, and are joined to the same Active Directory Domain and receiving the same Group Policies.
Windows Firewall is enabled on all client devices.
We have other Always on VPN RRAS servers for other clients built to exactly the same specification, and none of these experience this problem.
The server and the FW are both located in our private Datacentre.
We have also run SysLog but have not found any corresponding events that tie in with this.
Curious if anyone has any thoughts around what else this could be, or if anyone else has experienced this before.
Cheers.