When using perturbation theory the perturbation hamiltonian should be a relatively small correction to your problem. For example when solving atomic energy levels that are in eV the perturbation should only add a small contribution to this, like a couple of meV. In the case of a general shape of a nucleus, you even include the Coulomb interaction between the nucleus and electrons as a perturbation which is very unrealistic when you’re trying to determine the atomic levels because this is the main contributor (ignoring the keV-MeV of the nuclear states).