April 11, 2026
Paging the incident brain
Used Graphify to turn incidents into a queryable knowledge graph
Outages get a “memory map” and engineers are hyped—while skeptics clutch their pagers
TLDR: A new mashup turns Rootly incident data into a searchable “memory map,” promising faster fixes by surfacing past outages, what solved them, and what might break next. Commenters cheered the on-call superpowers while skeptics questioned accuracy, freshness, and whether this becomes just another pretty dashboard—still a big deal for teams fighting repeat failures.
Engineers are buzzing after a dev plugged Rootly (an incident tool) into Graphify to give outages a searchable memory. Instead of re-Googling every crisis, the tool builds a living map of what broke, who fixed it, and which services fail together. One ecstatic commenter, hamzmu, channeled Andrej Karpathy’s “LLM Wiki” idea and bragged that in “48 hours” we went from concept to “one command” incident brain. Translation for non-nerds: your past outages now act like a cheat sheet you can actually use.
Fans are calling it a power-up for on-call life—less panic, more receipts. They loved the visual goodies: a service heatmap to see what’s on fire, a who’s-on-call map to spot burnout risks, an alert-to-incident funnel to expose noisy alarms, and a follow-through tracker showing which fixes actually got done. The promise: faster answers to the three questions everyone screams at 3 a.m.—what broke before, what fixed it, what’s next.
But the drama is real. Skeptics side-eyed the “AI guessing game,” asking how reliable those connections are and whether a fancy graph just becomes another ignored dashboard. Privacy and maintenance questions lurked too: does this stay fresh, and who guards the data? Meanwhile, the meme crowd joked that outages “finally grew brain cells.” Love it or hate it, this thread made one thing clear: incident response might be getting its first real memory—and everyone has feelings about it.
Key Points
- •A Rootly-focused fork of Graphify connects to the Rootly API to collect incidents, alerts, teams, and services and builds a queryable knowledge graph.
- •Installation uses pip to add Graphify with the Rootly importer, and users must set a ROOTLY_API_KEY in a .env file.
- •The workflow includes a fetch/build step (graphify rootly) producing graph.html, GRAPH_REPORT.md, and graph.json.
- •An optional semantic analysis step runs via Claude Code or Codex to infer themes, patterns, and root cause relationships.
- •Users can explore heatmaps, on-call maps, alert-to-incident funnels, action item follow-through, and cross-service failure correlations, with in-browser filters for teams, severity, incident status, alerts, and time range.