By the concept of systems risk and resilience, this paper presents a technique based on probabilistic and stochastic modelling to gauge the adequacy of the supply—demand relation of rainwater tank harvesting in suburban Melbourne, Australia. A domestic rainwater harvesting system may be viewed from a technical (e.g. risk of demand not being met), economic (e.g. most economical tank capacity), or managerial (e.g. acceptable duration of time with empty tank) perspective. Rather than the traditional command-and-control approach to system selection and management, risk assessment provides a flexible way, in probabilistic term, to address the technical and economic perspectives, whereas resilience concept addresses the managerial perspective. This has the advantage of allowing a number of criteria for gauging the performance of alternative harvesting systems, as shown in an example by Monte Carlo simulation for a typical household rainwater tank system in Melbourne. This research provides a typical paradigm for analysis of systems of cluster or regional scales.