One good starting point will be the input power: if you have some estimate of the input power you can estimate the "nominal" current. The nominal current and power is estimated over some meaningful interval.

The output frequency also should be estimated to get a meaningful RF power output. This is not a conventional spark and a conventional spark plug is not appropriate.

Conventional spark plugs need a dead time to recover (consider a conventional discharge tube like a photo flash Xe tube discharge). Or, you can also consider a simple GM tube as a model.

I am afraid there is no "correct method" to get the "nominal current" for the discharge. You need to design your own devices- perhaps.