Skipping optimization in OptProf runs is useful to keep the OptProf pipeline running when no optimization input drops exist (perhaps because they expired, or it's the first run of the optprof pipeline).
But skipping optimization breaks the optprof pipeline's LKG feature that reuses older data when a test failure occurs so that the new drop has at least *something* rather than being empty, which breaks the main build pipeline. Lately, empty optimization input drops have broken our pipelines several times.

…ate/microbuild_staging' into libtemplateUpdate