To get the real info as to failure modes you would need to pry that info out of one or more of the manufacturers, possible one of the larger dealers. The sample size that any other hobbyist is exposed to is to small to draw a conclusion.

I have had servos fail where they drove hard over to one side. I have had them fail where they simply stopped working as if the power was turned off making them free running. In 45 years in R/C I can only remember 4 in flight servo failures. It is just not real common. 3 Of those were on Helios, the other one was an elevator servo on a 40 size 3D aircraft. I have had other failures but they were all done as a result of stupidity during setup. Maybe I have just been lucky to have such a small failure rate, maybe it is because I never use just enough servo for the job, I don't know.

Will anyone that has connections to a servo manufacturer or distributer chime in here with some real data? I would like to know the real answer also.