On a planet with more than 7 billion people, how do we identify the millions of drought-afflicted people who face a real threat of livelihood disruption or death without humanitarian assistance? Typically, these people are poor and heavily dependent on rainfed agriculture and livestock. Most live in Africa, Central America, or Southwest Asia. When the rains fail, incomes diminish while food prices increase, cutting off the poorest (most often women and children) from access to adequate nutrition. As seen in Ethiopia in 1984 and Somalia in 2011, food shortages can lead to famine. Yet these slow onset disasters also provide opportunities for effective intervention, as seen in Ethiopia in 2015 and Somalia in 2017. Since 1985, the US Agency for International Development's Famine Early Warning Systems Network (FEWS NET) has been providing evidence-based guidance for effective humanitarian relief efforts. FEWS NET depends on a Drought Early Warning System (DEWS) to help understand, monitor, model and predict food insecurity. Here we provide an overview of FEWS NET's DEWS using examples from recent climate extremes. While drought monitoring and prediction provides just one part of FEWS NET's monitoring system, it draws from many disciplines - remote sensing, climate prediction, agro-climatic monitoring, and hydrologic modeling. Here we describe FEWS NET's multi-agency multi-disciplinary DEWS and Food Security Outlooks. This DEWS uses diagnostic analyses to guide predictions. Mid-season droughts are monitored using multiple cutting-edge earth observing systems. Crop and hydrologic models can translate these observations into impacts. This information feeds into FEWS NET reports, helping to save lives by motivating and targeting timely humanitarian assistance.