Meta developed DrP, a root cause analysis platform that automates incident investigations for large-scale systems. Using DrP, over 300 teams at Meta run 50,000 analyses daily, achieving a 20-80% reduction in mean time to resolve (MTTR) incidents. The platform addresses the shortcomings of manual investigations, which often rely on outdated methods and can lead to prolonged downtimes. DrP features an expressive SDK that allows engineers to create analyzers, which codify investigation workflows. These analyzers are executed by a scalable backend that integrates with existing alerting and incident management tools, enabling automatic triggering during incidents. The post-processing system can perform automated actions based on investigation results, streamlining the resolution process. Engineers can easily create, test, and deploy analyzers, ensuring high-quality investigations that improve consistency and efficiency. DrP not only reduces MTTR but also enhances on-call productivity by cutting down the time engineers spend on repetitive tasks. Its ability to handle thousands of automated analyses daily makes it suitable for complex systems. Since its launch, DrP has evolved continually, improving its machine learning algorithms and expanding its integrations, ensuring it remains effective as organizational needs grow. Looking ahead, DrP aims to incorporate more AI capabilities, aligning with Meta’s broader AI4Ops vision for even more efficient incident resolution.