AI-Driven Monitoring and Operational Resilience

AI-driven monitoring, a multidisciplinary evolution of monitoring and monitoring automation, operates at the intersection of AIOps and observability, maximally interpreting observability data to achieve automatic hazard detection and response in critical infrastructures and systems. Such systems require operational resilience, which engenders the characteristics and capabilities that an organisation must possess to achieve safe and continuous operations, even when faced with unplanned events affecting the functioning of its products, services, processes, or business, thereby supporting reliable delivery of products and services within defined trust boundaries.

AI-driven monitoring has emerged as a critical function for maintaining operational resilience in business operations, enabling automatic detection of incidents as well as evaluation of their potential impact, in order to implement appropriate and timely responses. The design, implementation, and continuous, practical operation of AI-based, deep learning-based, or machine learning-based monitoring systems is extremely complex but can be decomposed into the following dependence structure of functionally distinct components: a stable data foundation providing the governance and quality of the data that creates the conditions for success; a reliable and accessible sensing, telemetry, and observability layer supporting optimised AI-based insights and interpretations of the data; a modelling and activity layer for the infrastructure in which the business services are hosted that enables modelling of sources of systemic risk; and a redundancy, failover, and recovery layer beyond the scope of this work that provides the additional capabilities necessary for operational continuity of services and processes.