Under Linux OpManager will look at the used memory and then display that, but this information includes the file cache which results in a value almost always in the high 90s and renders the counter useless.

What OpManager needs to do is take the value of used minus the buffers and cached values to return a value of how much memory is really tied up, since file cache can be dropped instantly and thus there's no reason not to use all memory not allocated to applications for the file cache.

e.g. 

             total       used       free     shared    buffers     cached
Mem:           502        494          8          0         88        343
-/+ buffers/cache:         62        439

OpManager would display a warning about this box as it is using 494MB of 502MB, however in reality the applications are only using 62MB and the rest has been allocated to the disk cache. Counting the disk cache as part of a memory count is not useful since it will always expand to use the available memory.