Now Earth's oblateness is 0.0033528590034 (0.33528590034%) based on measurements.
While some might not notice that from a big perspective, it's definitely 7 pixels larger than its polar diameter assuming it takes up 2160 pixels, so the polar diameter taking up 2160 px of the screen means the equatorial diameter would take up 2167 px of the screen, definitely visible to the naked eye (yet especially for people rotating/tilting the view), though it's definitely not much of a difference.
Still significantly important though.
Yet Earth's crust is 50 km thick in the continental areas and 5 km thick in the oceanic areas, so not exactly 2 dimensional(which would be a thickness of 0).
Now speaking of roche limit, we know it was a theory that large objects break up when too close to other large bodies due to tidal forces, and we witnessed that happen with the breakup of Shoemaker Levy-9 before its collision with Jupiter.
Now some other things (like the ISS for example don't break up by Earth's tidal forces, so not everything breaks up, therefore depends on consistency, mass, and also size, as many bigger bodies experience tidal forces distributed.
Yet Earth isn't a perfect elastic, but a nest of solids and liquids. with the crust being solid (and somewhat brittle) rock and the mantle somewhat solid, but slightly fluid, and the core being a near-spherical chunk.
Therefore within the roche limit, the crust would obviously break apart first before the mantle and core do, and it definitely wouldn't look like spaghetti (it wouldn't stretch a whole lot before you notice any fragmentation or glowing orange of melting and magma exposed, unlike putty, though definitely noticeable).