Process synthesis methodologies have traditionally focused on optimizing the economic criteria. However, owing to the complexity of real systems, external factors, such as expected or unexpected events or disturbances, negatively affect the performance of optimal networks. In this context, it becomes essential to consider the effect of such events on the performance of the network in the early design stages. This work presents a comprehensive review of the contributions related to this relevant topic, focusing on two widely utilized indicators: reliability and resilience. Reliability focuses on the probability of system failures due to disturbances, and resilience analyzes the behavior of the system after disturbances and its capacity to recover over time. Relevant contributions are reviewed, which present deterministic and stochastic methods to estimate these indicators, with a special focus on the design phase. Moreover, this work also presents a perspective on using machine learning methods on complex systems datasets as an emerging direction for enhancing the estimation of these properties. This contribution highlights recent advancements in this field and emphasizes the relevance of resilience and reliability as key metrics for developing safer processes with improved operability.