Rootly-graphify-importer transforms incident management by turning Rootly data into a queryable knowledge graph. This fork of graphify, inspired by Andrej Karpathy's LLM Wiki approach, builds persistent structured information that grows richer over time rather than rediscovering knowledge from scratch each query.

The tool operates in two phases: first collecting incidents, alerts, and teams via the Rootly API, then building nodes and connections between them. Users can explore service dependencies, identify coverage gaps, and detect cross-incident patterns through interactive visualizations that filter by team, severity, and time range.

Teams gain insights into recurring issues, single points of failure, and follow-through problems on action items. The graph-based approach reveals hidden infrastructure dependencies and shared-fate service groups through community detection, while confidence scores on inferred relationships ensure data reliability for critical incident analysis.