How much do eccentricities of different satellites in GSO vary?

It's improbable to get a payload into a perfect orbit, but obviously there is room for error, otherwise payloads in GSO would not work as well as they do. If the target eccentricity for a payload in GSO is 0, what is the margin of error? Have there been any egregious errors when maneuvering a payload to GSO?

Once in GSO, the eccentricity is perturbed by a few effects, particularly the pressure of sunlight (not the solar wind), on the satellite. It accelerates the satellite at 6PM local and decelerates the satellite at 6AM local, forcing the orbit to become elliptical. The eccentricity causes the satellite to oscillate east-west on a 24 hour cycle. Its control is part of the normal east-west stationkeeping. Generally the box is $\pm 0.05^\circ$ in longitude, but this can vary. A reasonable target eccentricity is $0.00028$ in this range.