The present invention provides for monitoring the health of a system. In one embodiment, a method includes storing benchmark data in a memory, storing a second plurality of system data indicative of system conditions during system operations, determining from the second plurality of system data that the system was operating at a steady state capacity level during at least one sample window of the system operations, associating the steady state capacity level of the at least one sample window with the benchmark data, retrieving first health data for parameters associated with a first health condition from the first plurality of system data, retrieving second health data for parameters associated with the first health condition from the second plurality of system data that were obtained during the at least one sample window, comparing the health data and determining if the second health data is indicative of a health condition.