A reliability assessment framework is presented for small unmanned aerial vehicles. The analysis considers several candidate architectures with different numbers of controllable surfaces and servos. It is assumed that a servo fault detection algorithm is available and affected by known rates of false alarms and missed detections. The aircraft flight envelope is analyzed to determine the fault levels for which the aircraft can still be flown at a trim point. For these "flyable" fault levels, it is assumed that the flight control law can be reconfigured to safely land the aircraft. Finally, the probability of catastrophic failure is estimated based on the histogram of (pre-fault) control command distributions, mean time between failure of the individual servos, and missed detection and false alarm rates. In applying the framework to assess the reliability of the candidate architectures, several interesting observations on design trade-offs are made.