In automaticCE, if the winning plan is chosen by CBR, we run the single solution through Multiplanning to measure its number of works and cache. In this case, we should avoid updating the server status metrics related to the MP exit conditions since it can be misleading.

One solution could be to introduce a specialized stage that handles measuring the works and caching and increments CBR specific metrics or we could find some way to plumb through that it is a CBR chosen plan and thus it should increment different metrics.