Incorrect RAM usage on PVE Node

[krom]

Describe the bug

top

top - 11:16:03 up 3 days, 15:46,  1 user,  load average: 4,92, 6,61, 5,82
Tasks: 376 total,   1 running, 375 sleeping,   0 stopped,   0 zombie
%Cpu(s): 50,0 us, 25,0 sy,  0,0 ni, 25,0 id,  0,0 wa,  0,0 hi,  0,0 si,  0,0 st 
MiB Mem :  31844,9 total,   2034,9 free,  20256,4 used,  10239,5 buff/cache     
MiB Swap:   8192,0 total,   8034,5 free,    157,5 used.  11588,6 avail Mem

For nodes with less load, the data is the same
To Reproduce
Steps to reproduce the behavior:

1. Load the node's RAM by more than 50%

Expected behavior
What you expected to happen.

Environment:

• Pulse Version: 4.24.0-rc.3
• Installation Type: Docker

Additional context
Historical data

[ozric70]

"memory":{
"total":16549875712,
"free":2467389440,
"used":6050492416
}

[masterchich-code]

после последнего обновления алерт спамит ТГ каждые 5 минут, не корректно начала отображаться рам память ноды (lxc), показывает нагрузку 90+%, хотя в реальности 50-60%

[krom]

This is PVE version 8.4.14
Version 9.0.11 is working fine

[Thormir84]

I have the same problem, but with some VMs; in my case, I have a VM that I use as a NAS based on OpenMediaVault; in Pulse, the machine shows the RAM almost full (96% - 15.3GB/16GB), complete with notifications; the fact is that from the OVM control panel, the RAM shows as being used at 61% (9.6GB/16GB) and the system works without problems.

Pulse is installed as an LXC on Proxmox, and is updated to version 4.25.0.