The statement (i) can be proven following Evan's book approach (Partial Differential Equations - page 38). I will assume that the function $f$ is continuous and bounded in $\mathbb{R}^2$.
In this post, it is shown that the Poisson kernel in the Half-space given by
$$ K(x,y) = \frac{2x_n}{n \alpha(n)} \frac{1}{|x-y|^n} \hspace{1 cm} (x \in \mathbb{R}^{n}_{+}, \hspace{0.3 cm} y \in \partial \mathbb{R}^{n}_{+})$$
(where $n\alpha(n)$ is the surface area of the unit sphere in $\mathbb{R}^n$) satisfies the relation
$$ \int_{\partial \mathbb{R}^{n}_{+}} K(x,y) dy = 1.$$
This can be used to prove the desired result. Indeed, if you fix $(x_0,y_0)$ in $\mathbb{R}^2$ and $\epsilon >0$, we may choose $\delta >0$ so that
\begin{equation}
\label{2}|f(x,y)-f(x_0,y_0)|< \epsilon \hspace{0.5 cm} \text{if} \hspace{0.5 cm} |(x,y)-(x_0,y_0)|< \delta. \hspace{1 cm } (1)
\end{equation}
If we take $|(x,y) - (x_0,y_0)|< \frac{\delta}{2}$, then
\begin{eqnarray*}
|u(x,y,z) - f(x_0,y_0)| &=& |\int_{\mathbb{R}^2} K(x,y,z,w)f(w) dw -f(x_0,y_0) | \\
&\leq& \int_{\mathbb{R}^2} K(x,y,z,w) |f(w)-f(x_0,y_0)| dw \\
&=& \int_{\mathbb{R}^2 \cap B((x_0,y_0),\delta))}K(x,y,z,w) |f(w)-f(x_0,y_0)| dw \\
&+& \int_{\mathbb{R}^2 - B((x_0,y_0),\delta))}K(x,y,z,w) |f(w)-f(x_0,y_0)| dw. \\
\end{eqnarray*}
Now, by (1) we have
$$ \int_{\mathbb{R}^2 \cap B((x_0,y_0),\delta))}K(x,y,z,w) |f(w)-f(x_0,y_0)| dw \leq \epsilon.$$
On the other hand, if $|(x,y)-(x_0,y_0)| < \frac{\delta}{2}$ and $|w-(x_0,y_0)|> \delta$,
we have
\begin{eqnarray}
|w-(x_0,y_0)| &\leq& |w-(x,y)|+|(x,y)-(x_0,y_0)|\\
&\leq & |w-(x,y)| + \frac{\delta}{2}\\
&<& |w-(x,y)|+ \frac{1}{2}|w-(x_0,y_0)|
\end{eqnarray}
Thus, we obtain $|w-(x,y)|> \frac{1}{2}|w-(x_0,y_0)|$ and therefore
\begin{eqnarray}
\int_{\mathbb{R}^2 - B((x_0,y_0),\delta))}K(x,y,z,w) |f(w)-f(x_0,y_0)| dw &\leq& 2||f||_{\infty} \int_{\mathbb{R}^2 - B((x_0,y_0),\delta))}K(x,y,z,w) dw \\
&\leq & C\frac{ z}{2 \alpha(2)} ||f||_{\infty} \int_{\mathbb{R}^2 - B((x_0,y_0),\delta))} \frac{dw}{|w-(x_0,y_0)|^{2}}
\end{eqnarray}
and this last integral tends to zero when $z \longrightarrow 0$, which shows that $u \to f(x_0,y_0)$ for any $(x_0,y_0) \in \mathbb{R}^2$.
I think that statement (ii) is not true. You may take, for instance, $f(x,y)=1$ and then,
$$u(x,y,z) = \int_{\mathbb{R}^2} K(x,y,z,w) f(w) = \int_{\mathbb{R}^2} K(x,y,z,w) = 1$$
and this function doesn't satisfy the required condition at the limit.