Systems and methods for monitoring health of systems and components of an oil rig are disclosed. Monitoring usage parameters allows for a probabilistic way to determine when rig equipment will fail. A failure probability curve based on past performance of the equipment can be used. Multiple usage metrics allow for increased accuracy and certainty of a time of failure for the equipment. Using a sufficiently high number of usage metrics allows the failure probability range to be very narrow and therefore the certainty of the prediction is high.