I was / am, aware that a resistor could be used to limit a spike, but using one like that is not efficient.
Somewhat changing the subject, an interesting device that I only worked with designing into a circuit one time in all the electronics I ever designed is a Varistor with a reverse voltage vs conduction curve. An industrial computer I worked on could do an incorrect boot if the power experienced a very short power disruption, resulting in the filter capacitor dropping low enough to not provide proper DC output from the power-supply, but for a brief time not low enough to trip the reset for a clean reboot start-up. The solution was to put a reverse slope Varistor of the proper voltage and power ratting across the main filter capacitor. When the voltage started to drop the reverse slope Varistor conducted a lot of current quickly discharging the filter capacitor, quickly bringing the 5 V supply low, resulting in tripping a circuit in the compute to cause a clean reboot when power came back on. It was a big deal, because otherwise the software was lost in sequence of running of that industrial computer that was controlling industrial processes.
A ups for only the computer would of been another fix, but insuring a clean reboot was good enough. Besides, the company using that industrial computer was not going to install a ups large enough to power everything it controlled.