Random RDS Session Crashes (Disconnections with black screen and only RDS banner) unable to re-connect until logged of by admin

Question

Problem: - Users report RDS session crashes usually but not always where the screen will go black leaving the only option to X out of the RDS banner. When trying to re-connect it will briefly allow login then immediately the session will drop and they will be back to the local desktop.

Environment:
RD Session Hosts (Windows Server 2016) x 4
RD Broker (Windows Server 2016)
Fslogix profile disks with redirected desktops (Windows Server 2022 - SSD array)
Office version - 365 MSO (Version 2206 Build 16.0.15330.20260)

Users - 60

Frequency of session crash - 1-3 per day on average. Sometimes same user is affected more than once. It occurs to people working from home and in the office.

Actions tried:

Removed Sentinel 1 EDR
Group policy for Session keep-alive
Disabled WIndows Firewall
Disabled Windows Defender
Fully Updated Hosts with Manfacturer drivers/firmware and Windows Updated
Removed RMM Remote Access Tool
Microsoft Incident - Got to requesting a clean boot on a server then stopped as that would render a production host unusable
Updated FSLogix

How long has it been happening - It has become more frequent since the profile disk was migrated servers in late April 2022 from a now redundant WS2016 server to a newer and faster WS2022 server. It had happened infrequently before then however.

I have tried investigating event logs on RD session HOst and Broker and while there are some errors such as ESENT 492, 416, 482. I haven't been able to track down the root cause. We get limite d opportunity to investigate as the users connect to the session for all their work.
Looking for further help

Answer 1

We ended up building a new set of four RD Session hosts (a sysprep image of one new server and duplicated three more times) to replace the affected 2016 Servers. It's been two months since doing this and daily occurances have stopped