It's interesting. Given the context (Krauss' larger argument) the point seems to be that among the uncountably many pieces of scientific data which the human race has carefully collected, prudently analyzed, ruthlessly abstracted, reduced, synthesized with other data, etc., we have never encountered anything remotely resembling a valid miracle that has withstood any degree of scrutiny.
Part of the problem may be the definition of the word. Just like "purpose" is meaningless (and even potentially cognitively harmful) when applied outside the context of human affairs, the word "miracle" doesn't have much meaning outside of a theological context. Absent that, in a secular context 'miracle' tends to simply indicate 'against all odds' or 'nearly impossible.'
I am not certain whether a serious answer to the question is really possible, given the problem that "miracle" has no effective transcendent meaning for science. There are things we do not know, sure — that does not make them miracles. The basic problem is that declaring something to be a miracle implicitly demands the priority of a particular theological interpretation of events, which is rarely satisfying to scientists and many philosophers.
The larger problem of why humanity would have need for the notion of miracles in the first place is probably better left to a response to a more specific question.