Sounds like your node might be crashing and rebooting?
The “Network.Socket.connect: <socket: 11>: does not exist (Connection refused)” error is returned when a query is made to the cli when the node is down.
Are you running as a systemd service? I have this problem when I accidentally run parallel services for my core node in both passive and active modes. Also, how many peers are in your topology (total valency count)? Can you determine whether or not your node is crashing every few minutes (keep in mind it takes about 5 minutes to come back up after restarting)?