In general, it's preferable to use 2D API's for 2D graphics. Yes, you can do 2D with OpenGL but it's significantly harder to get it right. The most obvious example is that in 2D we use pixels as coordinates and have no perspective to worry about, with 3D the your coordinates have to be carefully calculated every time the window or perspective changes.
It seems like you have never done 2d with openGl. Doing 2d in openGl is just a simple problem of using the correct transform matrix. Using pixel coordinates for your primitives is 100% possible and easy. The only difference between opengl in 2d and opengl in 3d is the transform.