(Most) Databases are arranged in tables, with a fixed and specific set of columns (item ID, glamour item ID, dye, spiritbond level etc.) using multiple dyes would entail adding new columns to this database and possibly modifying packet structures also if they rely on the definition of an item being a fixed number of bytes. Case-in-point: the Novus relic weapons from 2.x do not show a spiritbond meter once soulglazed - I strongly suspect this is because they reused the spiritbond counter to track how much light the weapon has accumulated.
This would require some way of flagging, for each item in the game, whether or not a character has ever owned it. There are thousands of wearable items in the game, so the glamour log underneath would be an enormous bitfield that grows substantially with every patch (I suspect this is also how the armoire works, though with far fewer items). Yoshi-P himself mentioned that they dould do something powered by achievements, in the same way that the Calamity Salvager checks achievements and quest flags when determining what you can buy from him. I thought that was a pretty good idea as it would require no additional flags at all. Definitely something is being worked on here though according to one of the live letters.
Because all the development effort for Yo-kai is already done. Bringing it (or any other previous event) back is a zero-effort change (they literally just have to set a flag), whereas doing a new event means new assets, dialogue, cutscenes etc. If there is not the time available for the people that make these things, then the choice is between bringing back an old event or having no event at all. Also, I'm not sure why you specified Monster Hunter, but in that specific case it's an IP not owned by SE, so there would have to be agreements negotiated (this would have happened for Yo-kai and GARO too), assuming Capcom would even be interested in a FFXIV crossover event.
Long story short, because we don't know how the game works internally, the community that asks for these things often vastly underestimates how much work they are asking for. Some of these things entail fundamental and far-reaching changes to the game, which would require a huge amount of testing and optimisation to maintain the quality to which we have grown accustomed, and all this while continuing to deliver new content at the same pace. It's not like the development team finished 4.1 a month after Stormblood released and is now just twiddling their thumbs and loafing around before they start on 4.2.