At what point do we start worrying that it's found a bug in Lean instead?
Right now.
Still possible, but Lean ships with formally verified kernels as of a few weeks ago.
In principle you can run Lean with different kernels and so there would have to be a bug in all of them for the proof to be faulty I believe.
Identically in all 4 independent checkers it ships with? Basically never.
The Lean kernel should be easy to verify by hand. We can ask the same question with humans though. What if a result that was used to prove some other result turns out to be wrong just because whoever reviewed it neglected a small detail?