I have a Ubuntu server (full desktop... don't ask) running 16.04.6 where DNS lookups stop working occasionally. Restarting the NetworkManager service resolves the issue, but only temporarily.
I'm able to detect the issue when a log collector stops receiving logs from this system. I access the server via SSH (yes, networking is working) and find that the log shipper service on the system is in a failed state. Attempting to start it fails because it can't resolve the destination server's name. I dig www.google.com and it returns "connection timed out; no servers could be reached". I restart the NetworkManager service, and now DNS lookups work and I can start the log shipper service, resolving the issue (until the next time it occurs).
I have previously commented out dns=dnsmasq in /etc/NetworkManager/NetworkManager.conf (How can I disable the DNS that Network Manager uses?) so that the system doesn't use dnsmasq as a local DNS proxy; however, I have this done on the other servers that perform the same role and don't have this issue on them.
My question is: why is this happening or what are the recommended steps to figure out the cause?
