With the notable exception of the kilogram, the SI base units are defined in a way as to make the units independently observable. In SI terminology, the coulomb is a derived unit. The reason is that measuring an ampere is a considerably easier (and more precise) task than is measuring a coulomb.
The reason is simply that current is easier to measure precisely than charge.
But it is still a very very difficult experiment if the accuracy is to be good enough. No one has -as far as I know- actually realized the Ampere this is way in quite a while (although I am sure there are a few old setups around).
In the "real world" the Ampere actually a derived unit, calculated from the realization of the Volt (Josephson effect) and the Ohm (quantum hall effect).
One of the big topics in metrology over the past few years has been to come up with a useful current standard, something that would allow us to define current in term of charges/second OR to realize the Ampere in terms of time, the latter would be useful since we can meaure time with higher accuracy than anything else (this is how the Volt is realzied. V=Kj*f, Kj being the Josephson constant adn f the frequency)