The "cannot connect to NoMachine monitor" error can be caused by a problem in the NFS configuration related to file locking. It was reproduced on RHEL 8, Debian 11.1, SuSE OpenSUSE 15.x.

Usually NFS has no problem with file locking, but some old kernels or NFS configurations might cause a file locking procedure to block indefinitely.  When this happens, the NoMachine process executing the file locking cannot answer to IPC requests, by consequently causing the NoMachine session to fail.

NFS v4 supports file locking out of the box, while NFS v3 requires a lock manager and a status monitor, usually supplied by NFS as lockd kernel service and statd daemon.

Troubleshooting

1) Check if the lock manager and status monitor are correctly working in the server NFS configuration. 
Daemons like rpc.lockd and rpc.statd should be running when the service is correctly started.

2) Check if a connectivity problem is preventing file locks from working correctly.
For example verify if lockd or statd ports listed in /etc/nfs.conf are being blocked by the firewall.

3) Verify if users' homes are mounted with the nolock flag, which explicitely disables file locking.
 

Possible workaround

As a workaround and a way to confirm that the behaviour is related to NFS, enable the UsersDirectoryPath key in /usr/NX/etc/node.cfg and set a different path for the creation of home directories of NoMachine users. If such path is in a non-NFS share, the problem will not occur.
 

Some more information is available in the official documentation of the Linux OS, for example for RHEL8:
https://access.redhat.com/documentation/it-it/red_hat_enterprise_linux/8/html/managing_file_systems/mounting-nfs-shares_managing-file-systems