So something like PRISM that targets everybody won't trigger a warning?
Keep in mind this will only work for non-court-gag-ordered instances. If the US subpoenas Apple about an individual they won't be allowed to notify them.
I have no idea how this applies to other countries.
I think this is more like: "We noticed unusual API usage and we don't have a gag order so whatever it is, it's not likely to be good"
Apple doesn’t need to know the source of the attack to issue the warning, and if the attacker is competent Apple likely wouldn’t know the source, such that a gag would not apply.
I don't see how Google could have been aware that this was happening, although they certainly could have known it was theoretically possible.