We study the error of the number of points of a lattice $L$ that belong to a rectangle, centred at $0$, whose axes are parallel to the coordinate axes, dilated by a factor $t$ and then translated by a vector $X \in \mathbb{R}^{2}$. When we consider the second order moment of the error relatively to $X \in \mathbb{R}^{2}/L$, one shows that, when $t$ is random and becomes large and when the error is normalized by a quantity which behaves, in the admissible case, as $\sqrt{\log(t)}$, it converges in distribution to an explicit positive constant. In the case of a typical lattice $L$, we show that this result still holds but the normalisation is more important, around $\log(t)$. We also show that when $L=\mathbb{Z}^{2}$, the error, when normalized by $t$, converges in distribution when $t$ is random and becomes large and we compute the moments of the limit distribution.