Leadshine EL8 interruptions

In a terminal
latency-histogram --nobase --sbinsize 1000 --show
leave it on, click on glxgears 5 to 10 times, watch some youtube, and take a screenshot after half an hour.

after a runtime of about half an hour with glxgears in different variants, the histogram according to the appendix was shown

That is perfectly OK.
Try in a terminal:
sudo lspci -v | grep Intel
or
sudo lspci -v | grep Ethernet
what do those return?

Which of these components can negatively influence the performance of the constellation Linuxcnc with Ethercat?

After a quarter of an hour of running the program, I spontaneously got the error message again, as shown in the attachment.Is this the latency time, can it be caused by the Ethercat network system or is another explanation plausible?

I have now achieved a situation where the system ran for 1.5 hours without interruption during test runs. To do this, I removed the EK1100 with several IOs and am only connected to the servo controllers with ethercat,

Too bad that I had to realize that ethercat with Linuxcnc does not run completely and consistently stable.

What do you think, what could be changed in the system so that this would be the case?