I can't get no crash dump

I've got a host that's mysteriously rebooting. There's no clue in the 
logs. It's not doing much - it's just a gateway. Just every few hours, 
it reboots. (It was a BSD 12 upgraded to 13, and it's been sitting there 
as a standby gateway for ages with no problem - ask it to do a bit of 
work and it gets in a strop).

Anyway, I'm trying to figure out how to get a crash dump, assuming 
there's a kernel panic leading to a reboot.

rc.conf:

dumpdev="AUTO"

# swapinfo -h
Device              Size     Used    Avail Capacity
/dev/ada0p2         8.0G       0B     8.0G     0%

# ls -ld /dev/dumpdev
lrwxr-xr-x  1 root  wheel  11 Oct  1 15:07 /dev/dumpdev -> /dev/ada0p2

# ls -al /crash
total 10
drwxr-x---   2 root  wheel   3 Nov 15  2022 .
drwxr-xr-x  24 root  wheel  24 Oct  1 15:07 ..
-rw-r--r--   1 root  wheel   5 Oct 23  2020 minfree   -   (contains 
"2048")

What am I missing here? Surely when it reboots I should be seeing 
vmcore.0 in there. Unfortunately this host is remote. Any other bright 
ideas about where I can find a clue what's upsetting it would be 
welcome.

Thanks, Frank.