I think you've got the essentials covered 🙂.
One other thing I can think of is the avg time / job execution. This maybe says something about the performance (and not capacity) of the server.
But to be honest, I would keep it simple to start with. For the purpose of load-balancing I would mainly look at total CPU usage of a server. Do some kind of round-robin mechanism, and run jobs on the server with the last load.