Abstract: Disclosed herein are system, method, and computer program product embodiments for detecting potentially unsafe wear levels in vehicle suspension systems. An embodiment operates by receiving acoustic signals of a vehicle suspension system, processing, the acoustic signals to generate training data, training a predictive learning algorithm to generate a predictive wear model and classifying the processed acoustic signals to indicate wear levels of one or more components of the vehicle suspension system. The trained predictive wear model subsequently predicts wear levels of the vehicle suspension system for notification to a cloud-based repository. Fleets of driverless vehicles may continuously monitor wear levels for timely maintenance before failure occurs.