$h(x)$ is monotone positive on $[0,\infty)$? It seems like it should be the case since $g$ must go down to $0$ faster than $h$, but I can't formalize the argument since $\frac g h$ can oscillate.
This came as a generalization of Find the $\lim_{n \to \infty}\frac{1}{n}\int_{0}^{n}xg(x)\mathrm{d}x$