Methodology

How BLACKOUT DESK computes risk scores and detects outages.

Data Sources

IODA (Internet Outage Detection & Analysis)

Georgia Tech's academic project monitoring global Internet connectivity via BGP, Active Probing, and Darknet data. Provides outage alerts and signal anomalies per country/ASN.

Cloudflare Radar (Optional)

When API token is configured, ingests traffic anomalies, confirmed outage annotations, and BGP hijack/leak events from Cloudflare's global network perspective.

Public Status Pages

Monitors Statuspage.io-compatible JSON endpoints for major platforms (Cloudflare, GitHub, Atlassian). Platform-level incidents contribute to the signal feed but don't directly trigger country-level blackout status.

Risk Score (0–100)

Computed per entity (country/ASN) every 60 seconds, based on weighted signal contributions:

SignalMax Weight
IODA active outage (last hour)+35
IODA recent outage (last 24h)+15
IODA signal anomalies+20
Radar confirmed outage (recent)+30
Radar traffic anomaly+15
Radar BGP events+15
Status page incidents+10

Status Thresholds

NORMAL (0–14): No anomalies detected
DEGRADED (15–39): Some anomalies, monitoring
PRE-BLACKOUT (40–64): Multiple corroborating signals or accelerating anomalies
BLACKOUT (65–100): Confirmed outage from multiple sources

Limitations

  • Scoring relies on public data sources — unmonitored regions may have blind spots.
  • IODA API response schemas may change; the worker handles this gracefully but may miss events temporarily.
  • Cloudflare Radar requires an API token; without it, coverage is reduced.
  • Status page data reflects platform-level issues, not necessarily country-level connectivity.
  • This tool is for awareness — it is not a substitute for official government communications.