From one of your comments to other answers:
I think the sunlight doesn't lose its blue that much on the way to the point where red scattering dominates.
Why wouldn't it be able to?
Rayleigh scattering has cross section proportional to $\lambda^{-4}$. The spectrum of light illuminating the volume scattering it gets multiplied by $\lambda^{-4}$, which indeed makes smaller wavelengths amplified more than larger ones. This indeed makes the scattered light, when seen immediately after scattering, bluer.
But on the other hand, this same scattering mechanism removes corresponding amount of power from the light that hasn't gotten scattered and continues propagating forwards. What we now have is this factor of the scattering cross section getting into the exponent of the Beer-Lambert law. In the limit of large distance $d$, the factor $\lambda^{-4}\exp(-d \lambda^{-4})$ has a reddening effect, not bluing. Here the distance you have to take into account is the sum of 1) distance traversed before scattering and 2) distance traversed to you after scattering.
Now, even if you take into account double scattering, which would give you a second factor of $\lambda^{-4}$, you'll still get $(\lambda^{-4})^2\exp(-d_\Sigma \lambda^{-4})$ where $d_\Sigma$ is the sum of distances traversed by the sunlight 1) before first scattering, 2) between first and second scatterings and 3) after second scattering until it got to you. This is even larger distance, and exponential function in the above mentioned expression, again, is much more selective by wavelength than the simple power in the factor it's multiplied by.
The result of this reddening, apart from the orange solar disk, is the phenomenon of Belt of Venus — the area of the sky on the opposite of the solar azimuth, which has purple-red color.
There's also a peculiarity of the Earth atmosphere that fools humans into thinking that the blue sky at day and blue sky at twilight have the same reason. Actually, if the atmosphere had no ozone, twilights also wouldn't even be as blue as they are. Beer-Lambert law would have much greater effect. On the Earth the twilight is blue because the sunlight, traveling in the upper parts of the atmosphere, gets absorbed in the red part of the spectrum by the Chappuis absorption band of ozone. If not the ozone layer, twilights would actually have sandy-brown color.