When designing approximation algorithms one sometimes solves a semidefinite program followed by a rounding step. An often used example to illustrate this is Max-Cut. (See e.g. Approximation Algorithms by Vijay Vazirani.)
Are there good educational sources or surveys going beyond the Max-Cut problem to explain more complex rounding algorithms and techniques used for their analysis? I'm thinking of cases when the vectors of the SDP-solution aren't distributed uniformly on a hypersphere, they have different lengths, or have other properties making the analysis harder.