Serve Stale on Fail

[geometricpanda]

Hey,
I'm loving this library as it's enabled me to centralise a redis cache across a number of pods :-) which is awesome, thank you

One challenge im wondering if this can solve is whether we can configure a longer stale time than revalidate time, and in the event that a revalidate call fails, then continue to use stale.

I'm struggling to get that to actually work in practice, have you got any advice?