Just a reminder: a young dragon is not a legendary creature, has no legendary actions, and has no lair actions.
Then it gets to die as easily as any other mook. If you intend for a bad guy to be a "boss" battle that lasts more than 2 rounds, then the boss needs stages and defenses, such as (1) ways to shed duration spell effects, (2) regain hit points, (3) change the battlefield environment and boost itself.
As soon as the PCs enter, the wizard casts suggestion on the half dragon and compels it to just up and leave. It fails. it leaves. Suggestion is RIDICULOUS.
I don't like the language either but I wouldn't have ruled it leaves. First, "sounds achievable" vs. "reasonable" is huge, but it basically must avoid even 1% risk of harm. Second, the example (leave the library) is situational. There might be harm involved with leaving the library forever (e.g. target is there to find a cure to its rare disease, so leaving now would be harmful), and there's certainly harm as below.
Charmed doesn't protect the rest of the party, only the caster. So, from the perspective of this dragon, I might rule that while the magic suggests I'd feel comfortable leaving my home in the care of my buddy for a bit, I think it'd be totally harmful if I turned my back on these other hooligans that I don't know who have weapons and armor and magic and look like thieves.
Suggestion doesn't allow you to tag on additional buddies or dictate they also are safe and best pals.
Anyway: how do you feel about save or suck spells in D&D 2024? Anything interesting, fun or frustrating to share?
They've been a part of the game forever. Throw in more low-level mooks, give bosses ways to shed effects. If you really don't like them, I'm trying a new mechanic in my next campaign:
anti stunlock, advantage on new save if you've failed an effect that disabled your ability to take actions. Doesn't affect ongoing effect saves, just new attempts to reapply the spell.