The problem with monitoring the line in current, is it's relation to motor current draw varies depending on the motor speed.

The motor will only see the full mains voltage when running at the rated speed. Any speed below the rated speed, the voltage applied is proportional to the speed i.e. 50% speed, 50% voltage.
What this means, if using a 2.2kw spindle rated at 220V with a 220V supply for example, is if you're running at maximum torque, at rated speed you'll be pulling 10A with the spindle ideally seeing 220V, and using 2.2KW. Now if you half the speed, the voltage drops to 110V, but the current remains constant at 10A, and you're only using 1.1KW. The supply voltage is still 220V, and due to the wonders of the switching in VFDs, the VFD will only draw 1.1KW/5A from the supply (that's assuming 100% efficiency - there will be some power loss).

Now from a spindle monitoring point, the supply current draw is useless, unless you're always going to combine it with the motor speed to get some kind of meaningful figure.


As for manufacturer's recommendations for supply fusing. They are based on worse case scenarios. Nearly all VFDs (and that includes servo drives, as they are essentially a VFD) have huge instantaneous switch on current surges. When first powered up, they appear as a short on the supply until their internal capacitor bank charges.
Then due to the switching frequency, you can get very high currents (why noise filters should be fitted), that can cause fuses to fail. Even though the average current will never exceed the rated power, there could be spikes many times the average current, which can be worsened if multiple drives share the same unfiltered supply, due to noise/frequency harmonics. In extreme cases, harmonics can cause wiring to burn out due to overheating, yet not trip any protection system.