In recent years, research on supply chain disruptions and internal failures has increased due to their significant impacts on extended downtime, reduced revenue, and diminished customer trust. Traditional reliability and risk management approaches primarily focus on failure prevention, whereas resilience extends beyond by incorporating prediction and adaptive strategies. Current methods for measuring resilience rely on classical reliability metrics, such as mean time between failures (MTBF) and mean time to repair (MTTR). While useful, these metrics fail to capture the multi-dimensional aspects of resilience, specifically the ability to prevent, predict, and adaptively respond to breakdowns. This limitation leaves production systems exposed to critical and unanticipated failures, making it difficult for practitioners to prioritize investments in predictive maintenance technologies or design effective resilience-enhancing strategies. To address this gap, this study develops a quantitative, data-driven resilience assessment framework tailored for production machinery. The framework employs Bayesian Networks (BNs) to model probabilistic dependencies among machine components, failure modes, and resilience strategies, enabling a holistic evaluation of machine resilience performance. A case study on a compound forming machine demonstrates the practical application of the methodology, highlighting its ability to identify critical failure modes and support decision-making for fault detection and human-in-the-loop assistance systems. By facilitating data-driven decisions and targeted retrofitting with emerging technologies, this approach enhances production resilience, reduces downtime, and provides valuable insights for both industrial practitioners and researchers.