Unexpected realtime delay error message

I got reports it threw Following Errors at some point- not right away.

Is it possible the NUC, being fanless, is silently overheating and the BIOS throttles it back and it can't keep up with the delays?  It was in an air-conditioned shop, probably 78F.

Is there any way to check what the status is, like processor temp?

OK, it did reoccur, and gave this while interpreting g-code.  This stopped it and put in the unpowered state, but they were able to Toggle Machine Power and get it going again.  At least that's the account I heard.   It looks like they were surfacing a board, I saw them running a similar thing when I came by.  I believe they said it happened when making an aggressive turn, but these are just regular right angle turns from like 200 ipm.  

When I did come by, they had been using the machine for awhile, and I felt the NUC box- it didn't seem especially warm.  Different job, but I don't think it's running hot and declocking to protect itself.
 
 

Can you run this command (in a terminal):

halcmd show param *.tmax

After LinuxCNC has been running for a long time
and post the results here (along with the CPU clock speed)

Hmm yes I had max accel at 50 and stepgen max accel at 60, which isn't 25% margin.

However, when I bumped up the axis to stepgen max accel to 70, then it immediately threw JOINT 0 FOLLOWING ERROR when trying to home.  That's why I had it set low.

What's causing the following error from stepgen being set too high?  With the prior install I had an optional parameter set that ran lower speeds but much, much higher max accel and stepgen max and never saw following errors.

You might try setting this higher:

setp hm2_[HOSTMOT2](BOARD).0.dpll.01.timer-us -50

say to -100 or -200 usec, because your system may have relatively
poor latencyso the DPLL baseline wander is more than -50 usec.

The following error limit is set low enough (0.0002 inches) that
a small timing error (~133 usec at 15 IPS) will trigger a following error
Too low of a DPLL sample time setting can trigger cause this timing error

Anything to see here?  It had been running awhile, but just slow linear passes to cut cabinets 

OK what's a better following error limit?