In my view, the single greatest issue in BG3 is how it handles spell/ability resource management: i.e., the long/short rest mechanic.
Short rests should be available whenever you're able to take a breather, not just once per day. The D&D 5th ed. rules as written place no limit on the numer of short rests a day and mandate a duration of 1 hour. Many DM's have house rules reducing this even further: Matt Mercer of Critical Role, for instance, allows his players to take short rests lasting only a few minutes.
Long rests, conversely, should be available only once per in-game day, and it is crucial that there be real, tangible consequences to the in-game passage of time, so players don't just blow all spell slots every fight. I can't stress this enough.
Now, while I personally abhor the resource recovery mechanisms of D&D 5th ed. and much prefer encounter-scope systems like the one in DOS2, BG3 is attempting to implement the former. Thus, class and overall game balance are impacted tremendously by how the implementation is carried out.
Assuming 5th level characters, four fights per in-game day, a short rest between each (i.e., four short rests), and finally a long rest:
* A fighter would be able to use 4 Action Surges (1 per fight), * A warlock would be able to use 8 spell slots (2 per fight), * A cleric would be ablel to use 4 (1st), 3 (2rd), and 2 (3rd) spell slots distributed as they saw fit across the day.
Assuming 5th level characters, four fights and unlimited long rests:
* A fighter would be able to use 4 Action Surges (1 per fight), * A warlock would be able to use 8 spell slots (2 per fight), * A cleric would be ablel to use 16 (1st), 12 (2rd), and 8 (3rd) spell slots.
Obviously, the second scenario is utterly imbalanced and would render content appropriate for fighters, warlocks, and other short rest-based classes utterly trivial for clerics, wizards, and other classes based on long-rest resource recovery.