The talk focused on observability tools, not experimental tools like micro-benchmarks. I probably should have made that more clear (I only mention experimental tools on slide 84).

I've actually been using fio a lot recently, especially since it can do non-uniform random distributions.