This is extremely valuable. Every time we get a problem which we are not able to reproduce, usually an extreme edge case, we end up getting our entire production DB replicated to get to the error.

I'll surely try this. Thanks for posting it here.