Session Timeout Connection Lost in Remote Desktop

Brass Contributor

Hey folks, have an odd one that we have not been able to remediate. Curious if anyone else has experienced this on a regular basis.


Using MSFT Remote Desktop app from home (on both corp managed devices but mostly personal devices) my team and I experience RDP Connection Lost errors routinely when using our WVD Spring Release personal desktops [screen cap below]. This happens in East US as well as in Mumbai. We will have this occur 5x a day sometimes, other days never.




Some of the things we believe we have eliminated from the list:


* We are using Azure VWAN and had some issues with it for a brief season, thought that could have been the cause, but those issues are resolved and it persists

* We had some who leveraged a FSlogix profile on an a separate Azure server (wanted to test FSlogix performance), moved to local profile on the session host, issue continues after disabling FSlogix

* We are all connected to a hard line inet connection, no wifi


Thanks for any insight! 


3 Replies


Hello, do you by chance have a NAT Gateway on the Virtual Network/subnet assigned to the WVD Desktops? 


We hit this bug with NAT Gateway and had pretty much the same symptoms as you.  if you do have NAT Gateway, remove it from the subnet.  then use a load balancer and NAT a high port to your desktop and block with Network security group as a work around to get the NAT but not break WVD.


I believe the fix is pending the product team.




@AndrewPhebus thank you for the response!


Ironic, we do have a NAT in place so all WVDs leave from the same public IP for ACL whitelisting. 


I will check into the recommendation you made and I will see if I can bug MSFT regarding this. MSFT actually recommended the NAT for this use case too.





My case with Microsoft on the issue is 120072525000034.  We were having the problem for about 4 months before a major NAT Gateway outage in the East US datacenter forced me to unbind the NAT Gateway from the virtual network subnet and then we implemented the Load balancer NAT to keep the same ACLS for the desktops. 


The Product team is apparently rolling out the fix for the NAT Gateway WVD stability issue, but i haven't seen the status that it is rolled out yet on my Ticket.