I want to show that $$ \lim_{n \to \infty} \left(1-\frac{n}{n^2} \right) \left(1-\frac{n}{n^2-1} \right) \cdot \ldots \cdot \left(1-\frac{n}{n^2-n+1} \right) = \lim_{n \to \infty} \prod_{k=0}^{n-1} \left(1-\frac{n}{n^2 - k} \right) = \mathrm{e}^{-1} $$
I know that $\lim_{n \to \infty} \left(1-\frac{1}{n} \right)^n = \text{e}^{-1}$ and so my attempt was to write
$$ \left(1-\frac{n}{n^2} \right) \left(1-\frac{n}{n^2-1} \right) \cdot \ldots \cdot \left(1-\frac{n}{n^2-n+1} \right)\\ = \underbrace{\left(1-\frac{1}{n} \right)^n}_{\to \text{e}^{-1}} \cdot \underbrace{\frac{1-\frac{n}{n^2-1}}{1-\frac{1}{n}}}_{\rightarrow 1} \cdot \ldots \cdot \underbrace{\frac{1-\frac{n}{n^2-n+1}}{1-\frac{1}{n}}}_{\rightarrow 1} $$ which I thought would solve my problem. But after a second look I see that splitting the limits in the product cannot be allowed. This would be the same nonsense as $$ \underbrace{\left(1-\frac{n}{n^2} \right)}_{\rightarrow 1} \underbrace{\left(1-\frac{n}{n^2-1} \right)}_{\rightarrow 1} \cdot \ldots \cdot \underbrace{\left(1-\frac{n}{n^2-n+1} \right)}_{\rightarrow 1}.$$
Maybe anybody can make the situation clear to me.
We have, by a simple comparison, $ \left( 1-\frac{n}{n^2-n} \right)^n \le \prod_{i=0}^{n-1} \left( 1-\frac{n}{n^2-i} \right) \le \left( 1-\frac{1}{n} \right)^n $. The limit of the left and right hand products as $ n \rightarrow \infty $ is $ \frac{1}{e} $, so by the Squeeze theorem, the limit of the middle product is also $ \frac{1}{e} $ as you wanted.