back
2 comments
I use fuzzing on top of a whole range of other test scenarios to expose tail risk (no matter how thin -> or especially for thin tails). It doesn't matter how long that fuzzer runs. What is more relevant is the quality of such a session which can by continuously improved by tweaking the fuzzer dictionary. This is what defines the limits of when it's no longer useful IMO. But because we aren't talking about minutes or hours but +/- days (or weeks), the "how long" is more of an academic question.
I believe you should never stop fuzzing.