Commentary;
For non-Modbus users, the apparently frequent modbus errors in the log are not especially alarming. It’s allowed for modbus devices to be too busy to immediately service requests, etc., resulting in a timeout at the openHAB end.
That’s why we have retries built in to the binding. Consistent “try 1/3 failed” means try 2 worked okay, so its just nuisance reporting and overhead.
It is sometimes possible to tune the binding parameters to reduce or eliminate the nuisance, that’s another topic really
Setting that aside, how extensive is your Modbus configuration? With default settings in place, Modbus binding is capable of pouring hundreds of updates per second into openHAB Items. This in turn can wallop persistence systems with default settings, and have serious performance costs.