Meta has transformed incident investigation into a scalable product called DrP. This platform automates debugging workflows by turning investigation expertise into reusable code. Engineers write 'analyzers' using an SDK, codifying steps like anomaly detection and data correlation.

These analyzers undergo rigorous code review, automated backtesting, and CI/CD deployment, ensuring reliability. DrP now handles 50,000 automated analyses daily across 300 teams, demonstrating its massive production scale. The system chains across service boundaries, connecting root cause analysis from API errors down to storage services without manual intervention. Results annotate alerts instantly, providing on-call engineers with diagnoses before they open dashboards.

Post-analysis closes the loop with automated remediation tasks and generates organizational learning from aggregated incident data.