Dear 庆甲 孟,
The error you describe is a known issue with Windows Server 2019 RDP services where the Remote Desktop Services stack (TermService) leaks resources over time, typically around the 90‑day mark, causing the RDP negotiation to fail. Restarting the service or the server clears the state, which is why it works again for another cycle. The root cause is not a port or firewall misconfiguration but a long‑running memory handle exhaustion in the RDP listener process.
The permanent solution is to apply the latest cumulative updates for Windows Server 2019, because Microsoft has addressed this in later builds (post‑2022 patches). If you are already fully patched and the issue persists, the supported workaround is to configure a scheduled restart of the Remote Desktop Services (TermService) before the 90‑day threshold, which prevents the listener from reaching the failure state. In environments where uptime is critical, Microsoft recommends monitoring the RDP listener health and automating service recycling rather than waiting for a manual restart.
If my answer is useful for you, please hit Accept the answer to support me.
Thank you,
HL.