Let:
\begin{align*}
&i = 1,...,m \quad \textrm{(index $i$ represent customers)} \\
&j = 1,...,n \quad \textrm{(index $j$ represent facilities)} \\
&y_{ij} = \dfrac{1}{n},\,\forall i,j \quad \textrm{represent the probability that customer $i$ chooses facility $j$} \\
\end{align*}
Each customer chooses 1 and only 1 facility independent of other customers. So here are a few questions that I couldn't solve:
$$P(\textrm{at least 1 facility gets no customer})\,=\,?$$ $$P(\textrm{1 and only 1 facility gets no customer})\,=\,?$$
Here is an attempt to solve the first problem:
$$P(\textrm{at least 1 facility gets no customer})=1-P(\textrm{all facilities get at least 1 customer})$$
It likely requires some sort of combinatoric analysis involving partitioning of the customers (elements) into facilities (buckets) but I don't think it is a trivial counting problem since all such partitions are not going to be equally likely to occur. I don't even know where to begin with in the first place, so any help would be much appreciated!
Edit 1:
With the help of Jimmy R., I found answer of the first question to be $1-\frac{\binom{m}{n}.n!.n^{m-n}}{n^m}$ and answer of the second question to be (both different from my comment below) $\frac{\binom{n}{1}.\binom{m}{n-1}.(n-1)!.(n-1)^{m-(n-1)}}{n^m}$ .
Hints: I assume that $m>n$, otherwise the problems are trivial. To tackle the first, think as follows: First assign one customer to each facility. So, "all facilities get at least 1 customer". Now, you need to distribute the $m-n$ remaining customers randomly to the $m$ facilities.
To tackle the second: just exclude one facility. You can do this in $n$ ways. Now, distribute all $m$ customers among the remaining $n-1$ facility, as if this one never existed.