I'll weigh in my thoughts on magic, and I'll do one point at a time.
Magic not being common in the world can easily be attributed to the fact that not everyone is inherently capable of it, and even those that are may never find out because they don't seek the training or... any number of factors. Those that are capable still may not be incredibly powerful because it does take
time and effort to learn magic.
As for the amount of magic one is capable of, it's likely gated by the same thing: how much time and effort they put into learning it, including their natural aptitude for it. There's definitely theory behind magic and it's inner workings, but for the most part it's bound to the will of the soul and so will only react to that. If one's soul isn't closely tied to the arcane, then they won't be able to do much with it. They can put in good effort to learn and understand it, but they will have a limit. Those with stronger bonds still have to actually learn and understand how to use it, or something like that.
The use of magic spells without pacing or moderation will result in some kind of exhaustion. The use of magic spells without greater focus, relevant utilities or reagents, and time, will probably result in less powerful or more inaccurate castings. For instance, one
can raise a wall against a pursuing army, but if one has not practiced this sort of spell, doesn't know the relevant foci or reagents (if any have been discovered), and doesn't take the time to ready the spell, it won't be as powerful, or it will be malformed, or perhaps the expenditure of magical force will just fizzle entirely if it goes badly enough. So it's between a number of factors there. The person's knowledge of that spell, the experience with that spell, the overall magical aptitude of the person, the power poured into it, the preparation done for it, and other relevant objects and stuff. From there it's up to player and GM discretion on what seems reasonable.
For instance, a bolt of fire could be a pretty easy spell to fling on command with a couple dramatic gestures and a couple command words spoken clearly to form the spell, maybe even less. An explosion may take more, requiring more movement, a somewhat longer command to shape the spell, and the thought to execute the spell as normal. To burst up a wall of flame may take the creation of an arcane circle, time to shape the spell with arcane command words, deep focus on the shaping of the spell in one's mind and where to summon it, and then the final execution which can still end up being quite taxing.
Spells cast from wands still need a basic aptitude for magic, but allow otherwise mundane casters to cast more powerful spells, only gated by the power of the materials made to use the wand itself. Therefore, magic for the mundane is quite expensive and inherently rare, so most wands will probably only be for lesser spells that one needs to cast frequently without thought or focus. There's probably some relationship between natural aptitude, expense on materials, and the type of spell... again, up to player and GM discretion.
The effect of a spell should be taken into account. For instance, a resurrection spell sounds like something blasé with the way you're putting it but can actually be a
game-changer when used on the right people. Something like resurrection should be
exceedingly difficult to pull off, require a lot of expenditure of all sorts of resources, and the person's actual willingness to
come back to life. So, if a spell has the power to become a tide turner, you should think about if it might have repurcussions if it goes poorly, or how much it takes to pull off. Mages like that will probably stick in the sidelines and wait for their time to shine when they can pull off their well prepared spells. Pulling out the ace up their sleeve that the enemy may not have prepared for, or they
may have. Fantasy medieval war may take on a new meta because of this and try to account for certain game changing spells from the enemy mages. But, I'm getting ahead of myself. Magic not being common in the world is no excuse not preparing for it if you know of its existence and capabilities. Although, some generals might
not prepare for it for those reasons, or just because they'
With that being how it is, it's still definitely viable for someone to do that, but spell casters in the field of combat like the ones we're playing will take on more hands-on roles with the war. A support mage might be a battle medic for one zone of the fight, watching over their soldiers with a keen eye, pulling back troops for recovery, bolstering others to push back, and protecting against the very same sort of tide turning stuff already spoken of in this conversation. Of course, even a battle medic will be fighting, so they get hands on experience in combat even if they are somewhere near the backline. Although, I figure they'd be closer to the
front even though it would be dangerous... high risk, high reward positioning, I guess.
I'd have to take option 2 between those two.
Three-five enemies in short succession at range is the Ranger's forte, and they could be much more versatile than a mage with limited range (i.e. sniping officers and key targets compared to Mages killing 3-5 frontline troops at once).
I tend to likened the heroes to positions in battle in my mind:
The Warriors are like infantry. Numerous, can take a beating, and are usually key forces and expendable simultaneously.
The Berserkers are like cavalry. They can deal some serious damage with momentum, but are easy targets when stopped.
The Rangers are... well, archers. Deal damage from afar, not so strong in close range.
Which leaves the Mages. The main battle positions have already been taken. All except...
Siege engines.
Glorious weapons of destruction and carnage, devastating armies and forts at range. Their drawbacks? They obviously can't directly fight, they are limited in number, and they require a lot of meat shields support to realize their potential.
Golems used as battering rams. Fireballs instead of catapults. But again, that's just what I imagined.
I do like the idea of more powerful spells requiring longer command times, and their are objects and artifacts that help shorten that time. Every second counts in a battle, and that trusty talisman can be the difference between life and death. I also like the idea the the GM personally checks on magic and makes sure it flies. It'll be extra work on the GM's part, but it'll allow people to personalize magic a bit more and not worry about accidentally breaking the war~
I like this, because it definitely shows the clear differences in the classes within war. Rangers/Archers and Mages are likely to fit the profile for
shock troopers as well, similar to berserkers but with some nuance maybe. A Ranger with their retinue of archers or warriors may lead an assault on an enemy group led out of position, where their arrows are most likely to cleanly hit enemies only and widdle numbers down until they can dash in for a quick execution, while certain mages may also be able to direct their more specific repertoire of combat spells to lock off these out-of-position troops, produce a bunch of shock-and-awe flashy evocations to obliterate enemies and cause confusion and rout enemies, and all manner of wondrous deeds. Singularly, Rangers can fit the bill for snipers, and that's REALLY good if there are high value targets to take down and you're specifically trained for that.
Now on the subject of Siege Engines/Operators... this is
definitely a conventional role for a mage, but it would also require the specific training for it. Siege Operators actually need to know
very well how to properly use their siege weapons and where to aim them to deal the most damage. The same can be said of a Siege Mage, which may be a subrole for the Mage rather than a hard and fast rule for them, just like other classes might have subroles. But, just like siege weapons are likely to have limited ammunition afforded to them in a siege, so it is the same with Siege Mages. They will need preparation, like the method of magic I originally blathered on about, and it will need
quite a bit of material. They are powerful, and they can do a
lot of damage if left alone in the backlines, but their influence will run out when they run dry on resources or tire themselves out.
So, therefore, it's a subrole. Mages are versatile by nature, so they could fit a number of good subroles, but people
really know the terror of mages through the fuck-off huge fireballs being lobbed at their castle walls and the golems bashing down their doors.
Anyway, that's my two cents on the subject.
The summary is, I like the idea of mages as siege operators, but I like having a little more versatility in the roles to allow for more specific kinds of mages, warriors, rangers, and berserkers. As for the magic system, I'd like for a little more variability in the way it works to allow for that kind of versatility, while also allowing for the specialization some mages will inevitably opt for and still leaving it as a reasonably balanced option for combat and utility.