They were produced at different times, with different imperatives of the monastic cultures that were reflected on what they needed at the time, and produced what we can call a “canon” if you like.
So the Pali Canon is an early collection, done even perhaps in an oral period. At that time the tantras and other such texts were not available—maybe not pronounced, maybe not compiled. The others, as collections, are much, much later. More diverse bodies of literature had evolved at that time.
And then within Tibet itself, and in China, too, I believe, there would be debates as to what to include as the authentic word of the Buddha. Because all of them define themselves as containing authentic words of the Buddha or disciples—there are different definitions of authenticity, but that is still the basic rule.
For the Tibetans, for example, they would only include texts that they believed had an Indian original it was translated from, therefore it was authentic.
Chinese also have a similar principle. We have old Chinese catalogs which discussed these questions. Sometimes a later cataloger would reject works that were accepted earlier on and say, “This is not an Indian compilation.” So the main idea was that it should be an authentic work of the Buddha, and it should go back to India if it was in translation.
But the decisions were made at different times, different places, so none of them were fixed really. Even the Pali Canon has some flexibility in it. Particularly the Tibetan and the Chinese evolved over centuries with debates, rejections of texts, taking in new texts and so on. So they are all evolving, changing.
Again, that’s one of the reasons the word “canon” is a little difficult, because the canon itself is an evolving idea.