User count bug?

[marte]

I started to notice a wrong user count which triggered my monitoring system with Proxmox V7.

After a little research there seems to be a bug with cleaning of users who where connected through a local WebIF:
Connecting via SSH (all PVE4):

10:56:12 up 2 days, 3 min,  1 user,  load average: 1.31, 1.29, 1.36
USER     TTY      FROM             LOGIN@   IDLE   JCPU   PCPU WHAT
marte    pts/0    10.100.51.5      10:56    3.00s  0.06s  0.02s w

Now connecting via web console from another host in the cluster (WebIF on PVE1 to Host PVE4):

10:56:39 up 2 days, 3 min,  2 users,  load average: 1.68, 1.38, 1.39
USER     TTY      FROM             LOGIN@   IDLE   JCPU   PCPU WHAT
marte    pts/0    10.100.51.5      10:56    0.00s  0.06s  0.02s w
root     pts/1    10.100.0.11      10:56    2.00s  0.03s  0.02s -bash

Closing Web console:

10:56:48 up 2 days, 4 min,  1 user,  load average: 1.54, 1.35, 1.38
USER     TTY     1 FROM             LOGIN@   IDLE   JCPU   PCPU WHAT
marte    pts/0    10.100.51.5      10:56    0.00s  0.06s  0.02s w

So - everything good ("1 user" again). But, if we connect tp PVE4 via WebIF on PVE4:

11:02:32 up 2 days, 9 min,  2 users,  load average: 1.63, 1.62, 1.49
USER     TTY      FROM             LOGIN@   IDLE   JCPU   PCPU WHAT
marte    pts/0    10.100.51.5      10:56    0.00s  0.11s  0.03s sshd: marte \[priv\]
root     pts/1    -                11:02    1.00s  0.02s  0.02s -bash

as expected. Now closing Web Console:

11:02:47 up 2 days, 10 min,  2 users,  load average: 1.33, 1.56, 1.47
USER     TTY      FROM             LOGIN@   IDLE   JCPU   PCPU WHAT
marte    pts/0    10.100.51.5      10:56    0.00s  0.12s  0.03s sshd: marte [priv]

aaaaand - something stays open ("2 users"). So there seems to be a bug that if you connect via WebIF on this host, the session is not closed successfully. When you connect from the WebIF of another host in the cluster, everything is fine. Can anybody confirm?

edit: clarification from which to which host