Next generation fleet wide asset monitoring solutions are incorporating machine failure prediction and prognostics technologies. These technologies build on signal processing of vibration time waveforms, process parameters, and operating conditions of the machine. For prognostics algorithms to work well, the signal processing algorithms need to be applied correctly and the results need to be reliable. This paper provides a survey of signal processing techniques as applied to specific machine component with a focus on the output and use with prognostics technologies. With properly organized outputs, prognostics algorithms transform the fleet condition and health management challenge into a deployable fleet health management solution. To arrive at the deployable fleet management solution, a systematic approach in the design of the prognostics system is preferable. This approach includes data and model driven failure patterns, sensory data connectivity from deployed assets, prognostics analytical applications, and advisory generation outputs which guide the asset owners and maintainers.