I don’t think you can have two inflow lines across the same cells. So I would probably make the secondary inflow its own boundary. If the peak flow overflows the main channel, then I would generally recommend extending the main boundary line across the flood plain. So the main boundary line would extend up to wherever you started the secondary inflow.
For the flow boundary, you have to enter a energy grade slope. The shallower you make this (smaller number closer to zero), the more the water will be spread out at the boundary. A bigger number will keep more of the flow in the channel.
The distribution of flow is not going to be perfectly accurate right at the boundary, but it should generally be ok as the flow moves downstream. You can adjust the slope up and down and see if it effects the model where you are trying to get good results.