This is needed in the get_indexer implementation for the new IntervalIndex behavior, as an overlapping IntervalIndex may return non-unique indices for a given query. Note that this equivalent to other indexes checking is_unique as part of get_indexer, which does not suffice for IntervalIndex for the aforementioned reason. Behavior and API design would be comparable to is_non_overlapping_monotonic.
My plan is to implement this within the IntervalTree to take advantage of some cached sorting that it performs, so it would not be implemented for IntervalArray (we don't build an IntervalTree there, as IntervalArray is mutable, and doesn't need indexing methods). This could be implemented for IntervalArray if need be but might not be quite as performant.