We always called this "margining".
Years ago, I had read a book about the Whirlwind
computer developed at MIT. In that book was the claim
that margining had been invented during that project,
the idea being to increase the voltage and make the
weak tubes fail so that overall reliability would be
improved after they were replaced.
I've wondered if that's really true or was the
technique in common practice and it just became
formallized at that time and entered into historical
lore.
Of course Whirlwind was also important for other
developments like core memory and employing a young
grad student by the name of Ken Olson.
BTW, I think the book was: Project Whirlwind: History
of a Pioneer Computer by Kent C. Redmond, Thomas M.
Smith Digital Press, 1980. Definitely a good read,
alas, my copy is miles away.
Regards, Jim