How Datadog defines service criticality at scale

To optimize deployment and reliability efforts at scale, Datadog developed a system to classify service criticality by combining multiple data signals, including LLM-analyzed postmortems, error traces, and traffic patterns. This centralized approach provides teams with consistent criticality levels and confidence scores via an API, enabling more efficient, automated decision-making across the organization.