Given positive integers, $n,m,r$, define $R((n,m);r)$ to be the least $N$ such that for any $r$-coloring $C:E(K_N)\to \{1,\dots,r\}$, there is some monochromatic subgraph with $n$ vertices and $m$ edges.
Surely this problem has been studied before. Is there a good reference for it, or any interesting conjectures here?
Edit:
I was aware of the fact that when $m\le \frac{1}{r}\binom{n}{2}$ that $R((n,m);r) = n$.
However, the problem still seems very interesting to me. For example, what can we say about the shape of $R((n,(1/2+\epsilon)\binom{n}{2});2)$? This should be at most (roughly) $n2^{2\epsilon n}$ by a modification of the Erdos-Szekeres argument for finding cliques. But if I use a random coloring, then I think we only get a lower bound of $2^{\Omega(\epsilon^2 n)}$ by appealing to a Chernoff bound.