Reliability and resiliency (R&R) definitions differ depending on the system under consideration. Generally, each engineering sector defines relevant R&R metrics pertinent to their system. While this can impede cross-disciplinary engineering projects as well as research, it is a necessary strategy to capture all the relevant system characteristics. This paper emphasizes the difficulties associated with defining the performance of smart microgrids, which are time-dependent systems and develops metrics and definitions for the design of electrical-mechanical systems that are valuable in assessing their performance, based on utility theory. A microgrid must not only anticipate load conditions but also tolerate partial failures or imbalance and remain functioning optimally. The microgrid operation during an outage differs from the normal grid operating condition, which optimizes as much as possible renewable integration and stability. Many of these failures happen infrequently but unexpectedly and therefore are hard to plan for. We discuss real life failure scenarios and show how the proposed definitions and metrics are beneficial.