One bones mesh vs Many Single meshes - how does it work?

I’ve seen the topic raised that sometimes its better to use a single mesh and have multiple bones for each object and animate each bones to reduce draw calls rahter than using a single mesh for each object. So how does this work? If we’re animating each bones individually won’t the paths be set?

Use case: I have a swarm of spaceships but since all animations are defined at compile time( correct?) won’t the paths my spaceships can travel be severly limited?

Can I break a bone off when its been destroyed? Might be better to make it invisible since object creation and destruction is costly.

you animate the bones fully through code. A bone is just another transform and thats all you need to move.

So the saving must come at load time then?