An average luxury car contains more than 50 sensors connected, to over 28 microprocessors, through multiple communication networks. What makes these complex machines diagnosable at a dealership, is the ability of sophisticated diagnostics algorithms. Besides use of diagnostics in service, diagnosing a failure is also key for functional safety and vehicle availability. Safety related diagnostic functions such as loss of Brake fluid and leaky fuel system detection are critical. Once a failure is detected, Vehicle availability functions extend vehicle operation, so that one could reach the dealership without being stranded. The number of failure modes in a car could far exceed tens of thousands, thereby identifying key failure modes that require diagnostics can be a challenge. Although regulations have done a great job of enforcing key diagnostic requirements through law, it is still essential to understand the science behind diagnostics development to provide high level of serviceability, Safety and Customer experience. This paper shares an approach for Comprehensive Diagnostics Methodology based on Sub-System Functional Failure Modes and Effects Analysis (FMEA), which is different than conventional way of developing diagnostics based on only Control System failure modes. Method to identify critical failure modes that require diagnostic algorithms, out of tens of thousands., based on Serviceability, Severity, Warranty cost, Regulatory requirements, Complexity of tool development, Failure reactions, etc., is explained through various examples. The paper also covers the economics of developing diagnostics On-board Vs Off-board to help with ever increasing memory requirements of new features.