I've always had trouble with this visualization too.
A better way of thinking about it might be to think of space as a 3d grid of springs. Matter pulls in the nearest springs, but this cause other springs in the grid to stretch out, creating tension.
Think of a relaxed spring stretched between two points. If you pull in just one part of the spring, the other side goes under tension.
/\/\/\/\/\ (a relaxed spring)
|||/-\_/-\ (a spring with the left parts of the coil pulled in, the right side is stretched out and thus under tension)
If you have two objects with mass in this "space" the "springs" between them act kind of like springs under tension. A simple model would be two weights on a table connected by a spring. If you pull them away from each other, the tension causes them to want to move towards each other. In the 3d grid of springs the same is true.
x - a mass
/-\_/-\_/|||x|||/-\_/-\_/ - a mass in space, notice space is pulled in around the mass, putting the rest of space in tension
x|||/-\_/-\_/|||x - two masses in space, notice space is pulled in around both masses, putting the space in between in tension
Nature wants to move towards an equilibrium, but it can't relax the springs around the mass, so to make this happen, the masses are pulled towards each other to remove the energy in the intervening space.
x|||/-\_/\_/|||x - starting to pull
x|||/-\/\_/|||x - more
x|||/-\/\|||x - more
x|||/\/|||x - more
and so on
except there's no physical spring in between, so while in the analogy, you'd finish with x||||||x, in the real world, the masses can keep moving towards each other and you end up with
x||||||x
x|||x - more
xx - the final resting state
We call this spring energy "potential energy" in physics, but it's caused by mass's necessary warping of space-time.
I think this is easier and more intuitive to visualize in true 3d. The warped 2d fabric is very hard to visualize in 3d and the way it functions (an outside force of gravity pulls objects down the slope towards the notional masses) isn't quite accurate.