Let $A = \begin{bmatrix}
1 & cos\theta\\
cos\theta & 1
\end{bmatrix}$ and the eigenvalues be represented by $\lambda$,then eigenvalues can be determined with the help of below property :
$det(A-\lambda I = 0)$
We get $2$ values of $\lambda$ as the matrix size is $2$x$2$.
$\lambda_1 = 1 + cos\theta\\\lambda_2 = 1 - cos\theta$
Now, let $\textbf{x} = \begin{bmatrix}
x_1 \\
x_2
\end{bmatrix}$ denotes the eigenvector, then by the property of eigenvector,
$\begin{array}\\
& A\textbf{x} &= \lambda \textbf{x} \\
\implies&(A - \lambda I)\textbf{x} &= 0
\end{array}$
On solving for $\lambda_1$, we get
$\begin{bmatrix}
-cos\theta & cos\theta\\
cos\theta & -cos\theta
\end{bmatrix}
\begin{bmatrix}
x_1 \\
x_2
\end{bmatrix} = 0
$
which simplifies to $x_1 = x_2$.
As an eigenvector merely refers to the ratio between the two (or more) quantities, we say that for $\lambda_1, \textbf{x} = \begin{bmatrix}
1 \\1
\end{bmatrix} $.
We usually write eigenvector in normalized form, i.e. we divide each element of the eigenvector by $\sqrt{x_1^2+x_2^2\ldots x_n^2}$ for an eigenvector of size $n\text{x}1$.
So, the normalized eigenvector is $\begin{bmatrix}
\frac{1}{\sqrt{2}}\\\frac{1}{\sqrt{2}}
\end{bmatrix}$.
Similarly, the eigenvector for the other eigenvalue can also be determined.
For more details, please see this.