I have multiple Ubuntu 18.04 LTS Server installations on the same vSphere 6.7 host. One runs my DHCP and DNS server and has done so reliably for years, all the way from 14.04 to 18.04.
However, one of the servers is suddenly getting moments (multiple times per day) where it can't use DNS to resolve addresses. If you ping something from the command line, you get 'Temporary failure in name resolution'. This can last from 30 minutes to 12 hours. Eventually - with no changes on my part - it will start working again, only to fail again at some point. Resetting the host tends to resolve the issue for a while too.
This server was recently installed (about 2 weeks ago) 18.04 from the ISO and contains all the latest patches. I believe that because it is from the default that means it is using resolved and not interfaces, but I don't know for sure as I haven't followed how Ubuntu changed in that area. I do know that this host gets its address from DHCP and on a reboot, it reliably picks up not only its address but the DNS server list just fine (systemd-resolve --status displays all the DNS server IPs correctly).
What I don't know is how to debug this? What should I be looking for?
It is super challenging to learn when it happens - I notice it much later when I realize that its rsync jobs to update the other servers with weather data isn't working and I now know from experience that this means name resolution isn't working. Is there more debug info than what is in syslog? Any thoughts on how I can pin down when it starts happening to see if there are relevant syslog entries? Also, is there better debug info from whatever is doing name resolution (sockets?) when I attempt it and it fails?
Should I be looking for anything on the DNS server itself? I am the sysadmin for it so I can do whatever I need to.