Shifting ecosystem disturbance patterns due to climate change (e.g. storms, droughts, wildfires) or direct human interference (e.g. harvests, nutrient loading) highlight the importance of quantifying and strengthening the resilience of desired ecological regimes. Although existing metrics capture resilience to isolated shocks, gradual parameter changes, and continuous noise, quantifying resilience to repeated, discrete disturbances requires novel analytical tools. Here we introduce a flow-kick framework that quantifies resilience to disturbances explicitly in terms of their magnitude and frequency. We present a resilience boundary between disturbances that cause either escape from a basin of attraction or stabilization within it, and use the resilience boundary to build resilience metrics tailored to repeated, discrete disturbances. The flow-kick model suggests that the distance-to-threshold resilience metric overestimates resilience in the context of repeated disturbances. It also reveals counterintuitive triggers for regime shifts, such as increasing recovery times between disturbances, or increasing disturbance magnitude and recovery times proportionately.