I'm fresh to homological algebra, and our textbook is Weibel's.
When I was showing the Exercise 1.2.2 of Weibel's book, I'm in the following "path" so that the notations of kernel, monic and monomorphism are the same
(1) Every kernel is monic;
(2) Every monic morphism in $R$-$\mathbf{Mod}$ is a monomorphism;
(3) Every monomorphism in $R$-$\mathbf{Mod}$ is a kernel of a morphism.
When I'm doing (3) the monomorphism $f:A\to B$, I first define the quotient map $\pi: B\to B/\mathrm{im}{f}$ and show that the inclusion $i:\mathrm{im}{f}\to B$ is a kernel of $\pi$, and then use the universal property of $i$ to show the universal property of $f:A\to B$. And in my argument, I use the monomorphism property to show that there is a bijection from $A$ to $\mathrm{im}f$, and thus the domain of $f$ and $i$ are "the same" in some sense, and their codomain are both $\mathrm{im}f$, and it seems that this "the same" preserves the composition of morphisms.
Here is my question, are all kernels of a morphism the same in some sense?
I've read that the kernel and the usual $\ker{f}=\{a\in A: f(a)=0\}$ are the same, and the latter one is unique obviously, then all the kernel of a map in $R$-$\mathbf{Mod}$ would be the same in some sense. But in what sense?
The categorical definition of kernel is a special case of limit, and limits are unique up to isomorphism.
Now this means that, if $k_1:K_1\to A$ and $k_2:K_2\to A$ both satisfy the universal property of being the kernel of a (nonnecessarily injective) morphism $f:A\to B$, then there is a (unique) isomorphism $u:K_1\to K_2$ such that $k_1=k_2\circ u$.
On the other hand, if $k_2$ is a kernel of $f$, and $u:K_1\to K_2$ is an isomorphism, then it is easy to see that $k_2\circ u$ is again a kernel of $f$.