First off: I'm not an expert in order theory, so some of my terms might be off; correct them if you wish.
Let me call a subset $A$ of a lattice $(S,\le)$ upwards convex (not sure if that's actually good terminology) if it holds that for all $a,b\in A$ and all $x\in S$ with $a\le x\le a\vee b$ (where $\vee$ is join), also $x\in A$.
Now let $M$ be a finite set with $n$ elements and $P\subseteq 2^M$ any subset of its power set. $P$ is (of course) not necessarily upwards convex, but it can be covered by upwards convex sets: there exist subsets $P_1,\dots,P_k\subseteq P$ such that all $P_i$ are upwards convex and $P=P_1\cup\dots\cup P_k$. (One such covering is given by $P_1,\dots,P_k$ being the single-element subsets of $P$, proving the above existence claim.)
Say that $P$ is $k$-coverable if there exists a covering $P_1,\dots,P_k$ by upwards convex sets. Two questions:
What is the minimum $k$, as a function of $n$, such that any $P\subseteq 2^M$ is $k$-coverable?
What is the computational complexity of determining, given $P\subseteq 2^M$ and $k$, whether $P$ is $k$-coverable?
As a partial answer to (1), I can show that $k\ge n(n+1)/2 + 1$ using a recursive argument.