The Internet exposes us to cyberthreats attacking information, services, and the Internet infrastructure itself. Such attacks are typically detected in a reactive fashion. The downside of this approach is that alerts of an attack are issued as it is happening. In this article, we advocate that the security community could benefit by complementing traditional reactive solutions with a proactive threat detection approach, as this would enable us to provide early warnings by analyzing and detecting threat indicators in actively collected data. By describing three use cases from the DNS domain, we highlight the strengths and limitations of proactive threat detection and discuss how we could integrate those with existing solutions.