Panda Sigma unexpected Crashes (Ubuntu)

In March 2025 we bought Panda Sigma 32GB RAM, unfortunatelly experiencing unexpected Crashes (rather freezes, very probably NOT caused by the OS).  

How the crashes manifest themselves:
Display Screen connected to the panda's HDMI port goes black, 
ICMP ping to machine IP addr does not respond, 
hitting the USB keyboard does not help, 
tried to reconnect the keybord, no luck with waking up the machine, 
tried to reconnect the keyboard to USB C - same results, 
cannot light up even the keyboard numlock LED, 
the panda's onboard blue LED diode lids as usually, 
panda's FAN rotates as usually, 
need to manually power the machine off an on. 
After that the machine runs as usual.

Machine config description:
Panda Sigma 32GB RAM using original power adapter, 
Flashed with LP-BS-WTADLC1R210-32G-A-ON (Auto Power-On BIOS)
1x WD Red SN700 NVMe 1TB, BTRFS Raid1    in the M.2 M 2280 PCIe 3.0 slot
1x WD Red SN700 NVMe 1TB, BTRFS Raid1    in the M.2 M 2280 PCIe 4.0 slot
Normal HDMI display screen connected for Debug purposes, 
Normal keyboard into front USB, 
Ubuntu 24.04.2 LTS SERVER, 
Microservices installed like MariaDB, Grafana etc, some Python API's. Nothing special.

After the Crash (freeze), Ubuntu's journalctl doesn not show anything unusual, no OM's, no kernel panics, no attempts to go to sleep, nothing unusual, 
same for /var/log/dmesg and /var/log/kern.log - no suspicious information.
First i suspected that the panda's bios power mangement forces OS to some sleep, but that was not confirmed. 
I have successfully engaged Serial Console for debugging, but no poisonous output has been found there, no kernel panic, not a single extra character, just regular info.

First month the Sigma Crashed like every day or every fifth day, mostly at night. 
Btw no production or system service or crontab action was schedulled to run around the crash time....
After i disabled the in bios In-Band ECC Support, it looked like the crashes stopped, but after cca 25 days later the machine crashed again.

Bios settings:
- RC6 (Render Standby) DISABLED. 
- Enable Hibernation - DISABLED
- ACPI Sleep State - Suspend Disabled
- D3 Cold Disabled
- D3 Hot Disabled

cat /proc/cmdline
BOOT_IMAGE=/boot/vmlinuz-6.8.12 root=UUID=6d738a37-9aea-4be5-afcf-6132f214df09 ro ipv6.disable=1 rootflags=degraded systemd.log_level=debug usbcore.autosuspend=-1 console=ttyS1,115200n8 console=tty1 nvme_core.default_ps_max_latency_us=0 loglevel=7 debug earlyprintk=serial,ttyS1,115200n8 crashkernel=512M
I tried everything in the cmdline, like pcie_aspm=off, pcie_port_pm=off pci=realloc intel_iommu=off iommu=off pci=realloc
but nothing helped.

So, after two months of doubtful operation i am desperately slowly preparing myself to shut down and remove the Sigma from production operation.
While being not happy with such situation, because sigma is very, very fast beast. 
..unless a miracle happens, e.g. learning something salutary here, at the forum.

So, according information i gathered to the day, i am convinced that it must be something hardware related as the freeze happens immediatelly with absolutelly no traces. 
 

Please give me some hints what can be happening.

P.S: Today i have desperatelly reflashed bios, resetting the variables that i changed overtime during desperate effort "make it better".