stagenet

The problem came back after a while. After thorough testing, I noticed that the switch uplink ports was in a 100/100 Mbit/s. I added a policy to the port to disable the port if the switch found a loop. The port got immediately blocked. I changed the switch and everything seemed to work. Therefore I expected the cable, the original switch or the port itself.
 
After some trial and error I verified that everything worked with another switch. My original switch had been on for 90 days so I rebooted it. After that everything started to work perfectly. Took me a while to get there because I was staring myself blind on the logs on Unraid. However, now everything works so I am happy.  

In the syslog.txt in the diagnostic I found this:
 
Feb 11 14:10:55 AG5-NAS-01 kernel: igc 0000:09:00.0 eth0: PCIe link lost, device now detached  
Which led me to this:
 
Almost the same hardware. Trying to add this.
 

 
Will get back with results.