Intermittently losing IPv4 connectivity on WAN

Question

For a couple of weeks now, I've been encountering intermittent loss of WAN connectivity. Looking in the System log, I see the same sequence of events every time: 
 /var/log/system.log:2018:07:20-01:40:32 astaro dhclient: Killed old client process 
 /var/log/system.log:2018:07:20-01:40:33 astaro dhclient: DHCPREQUEST on eth3 to 255.255.255.255 port 67 
 /var/log/system.log:2018:07:20-01:40:36 astaro dhclient: DHCPREQUEST on eth3 to 255.255.255.255 port 67 
 /var/log/system.log:2018:07:20-01:40:36 astaro dhclient: DHCPREQUEST on eth3 to 255.255.255.255 port 67 
 /var/log/system.log:2018:07:20-01:40:37 astaro dhclient: DHCPACK from 69.134.151.22 
 /var/log/system.log:2018:07:20-01:40:37 astaro dhclient: bound to 173.174.xxx.xxx -- renewal in 27171 seconds. 
 where the address in the last line is the same address I had before the problem. The timestamps coincide exactly with the "Internet uplink is down" email notifications that the UTM sends me when this happens, but despite the apparent successful "renewal" of my WAN address, I am no longer able to contact the outside world via IPv4. 
 However, it appears that connectivity is not lost altogether... 
 
 I can still access the web interface of my cable modem at 192.68.100.1 (I have 192.168.100.2/24 defined as an additional address on my WAN port), and 
 IPv6 connectivity continues unabated (which would be great if IPv6 uptake was more universal, but unfortunately things like my company VPN are still IPv4 only). 
 
 Power cycling the modem (an Arris SB6183) usually clears the problem for another couple of days, though a couple of days ago I had to reboot the UTM as well. 
 Any ideas what might be causing this, or where I might look for more background information? As a Free Home User, I don't have access to Sophos support.