I've had a lot of instability issues with my server, where the only fix is physical forced reboot. It seems to mostly been when writing, or heavy disk load, but seemingly not always. I have not at all figured out what is wrong with it. 
The parity checks have always come out fine. 

But when I installed scrutiny to check the S.M.A.R.T status of the disks it shows the parity disk has 36% Reallocated Sectors Count. 

"Count of reallocated sectors. The raw value represents a count of the bad sectors that have been found and remapped. Thus, the higher the attribute value, the more sectors the drive has had to reallocate. This value is primarily used as a metric of the life expectancy of the drive; a drive which has had any reallocations at all is significantly more likely to fail in the immediate months"

It seems obvious I need to replace the parity drive, and will do so.
But what I'm more curious about is if this can be the reason for the instability issue?

When I write a lot to the disk, unraid write it to the disk and to the parity disk, the parity disk get to a bad sector and have to remap it, and then another one, and then another one simultaneously, and somewhere this crashes unraid? 

scrutiny also semi-crashed unraid, the gui or cli barely worked at all, managed to open 1 page in 10 minutes, and run one show command in the same period but failed to run any executive command (like turning off the scrutiny docker), while all dockers, scrutiny included, was working fine, so might be something messed up with a disk's S.M.A.R.T and unraids ability to handle this messed-up-ness. I had to give up and physically reboot and unraid was flowing fine again with scrutiny turned off. 

1 minute ago, tshorts said:

and somewhere this crashes unraid? 

It should not, unless the disk is timing out, that can give the appearance of crash, diags/syslog would help

7 minutes ago, JorgeB said:

It should not, unless the disk is timing out, that can give the appearance of crash, diags/syslog would help

Unfortunate.
I tried the built in smart self-test that went Completed without error, perhaps different the disk isn't so bad as scrutiny say it is, and 36% doesn't mean the way I interpret it. 
Regarding syslog I've always only managed to see log from the last boot, not before that, as in not what's logged during the crash. Can I make it persistent somehow? 

You can enable the syslog server.

You are likely to get better informed feedback if you attach your system’s diagnostics zip file to your next post in this thread. It is always a good idea when asking questions to supply your diagnostics so we can see details of your system, how you have things configured, and the current syslog.   Among other things this  would allow us to see the SMART status of your drives.

I just want to announce, in case other experience similar issues in the future. 

That I've replaced the parity disk with the relocated sectors with a new disk, and all instability issues have ceased. 

It's now been up for a week. 

While the longest before was like 2 days before hard reboot was needed. Sometimes several times in a day.