Let $b > 0$ be a fixed real constant and take $z_a = ia + b$. Then $\mathfrak{Re}(z_a) = b$ for all $a\in\mathbb{R}$. I have managed to confuse myself by reasoning that $\lim_{a\to\infty}\mathfrak{Re}(\sqrt{z_a}) = \infty$ in the following way:
Write $w_a = \sqrt{z_a} = \sqrt{ia + b}$. What is the asymptotics of $\mathfrak{Re}(w_a)$ as $a\to \pm\infty$? $z_a$ form the vertical line intersecting the real line at $b$ and $z_a$'s principal argument clearly satisfies
$$\mathrm{Arg}(z_a)\in\left(\frac{-\pi}{2},\frac{\pi}{2}\right)$$
for all $a$. This would then lead me to believe that $\mathfrak{Re}(w_a)$ is finite for all $a$, but
$$w_a = \sqrt{|w_a|}e^{i\mathrm{Arg}(w_a)} = \sqrt{|ia + b|}e^{i\mathrm{Arg}(z_a)/2}$$
and $|ai + b| = \sqrt{a^2 + b^2}$. Thus
$$w_a = \left(a^2 + b^2\right)^{1/4}e^{i\mathrm{Arg}(w_a)/2}$$
so that
$$\mathfrak{Re}(w_a) = \left(a^2 + b^2\right)^{1/4}\cos\left(\mathrm{Arg}(z_a)/2\right)$$
Since $\mathrm{Arg}(z_a)/2\in \left(\frac{-\pi}{4},\frac{\pi}{4}\right)$ for all $a$, then $\cos\left(\mathrm{Arg}(z_a)/2\right) > 2^{-1/2}$ and so
$$\lim_{a\to\infty}\mathfrak{Re}(w_a) = \infty$$
which seems nonsensical. What is going on?
Here is the graph of $\operatorname{Re}\big(\sqrt{ia+1}\big)$. Yes, it does go to $\infty$ as $a \to \infty$.
Why do you say it is nonsensical? In fact $$ \lim_{a\to\infty} \frac{\operatorname{Re}\big(\sqrt{ia+1}\big)}{\sqrt{a}} = \frac{1}{\sqrt2} $$