The degradation in electrical output of solar arrays on Mars landers and rovers is reviewed. A loss of 0.2% per Sol is typical, although observed rates of decrease in 'dust factor' vary between 0.05% and 2% per Sol. 0.2%/Sol has been observed throughout the first 800 Sols of the ongoing InSight mission, as well as the shorter Mars Pathfinder and Phoenix missions. This rate was also evident for much of the Spirit and Opportunity missions, but the degradation there was episodically reversed by cleaning events due to dust devils and gusts. The enduring success of those rover missions may have given an impression of the long-term viability of solar power on the Martian surface that is not globally-applicable: the occurrence of cleaning events with an operationally-useful frequency seems contingent upon local meteorological circumstances. The conditions for significant cleaning events have apparently not been realized at the InSight landing site, where, notably, dust devils have not been detected in imaging. Optical obscuration by dust deposition and removal has also been observed by ultraviolet sensors on Curiosity, with a similar (but slightly higher) degradation rate. The observations are compared with global circulation model (GCM) results: these predict a geographically somewhat uniform dust deposition rate, while there is some indication that the locations where cleaning events were more frequent may be associated with weaker background winds and a deeper planetary boundary layer. The conventional Dust Devil Activity metric in GCMs does not effectively predict the different dust histories.