How rxmna8502/sentinel-ni works
Sentinel / NI is an out-of-band network-security forecasting platform. Its main operator journey enrolls a remote sensor, collects and submits aggregated telemetry, builds ten-second network states, forecasts five horizons, then presents source intelligence and mitigation recommendations in the dashboard. The graph includes the agent-to-central and browser-to-central boundaries, sensor management, and the forecasting and recommendation stages. Some API/runtime and collector wiring is unsampled, so only documented or excerpt-supported links are shown.
Main entry points: live_runtime.py Sensor runtime · page.tsx Dashboard UI · app.py Central API · service.py Agent service
Components 22
Telemetry collection
- Agent service
src/agent/service.py - Telemetry client
src/agent/client.py - Telemetry adapters
src/telemetry
Central platform
- Central API
src/api/app.py - Sensor runtime
src/api/live_runtime.py - Sensor management
src/sensors/manager.py - Sensor registry
src/sensors/registry.py
Forecast and response
- State aggregation
src/streaming/state_aggregator.py - State history
src/streaming/state_buffer.py - Forecast inference
src/forecasting/inference.py - Forecast explanation
src/forecasting/explanation.py - Source intelligence
src/streaming/source_forecast.py - Mitigation policy
src/evaluation/mitigation_policy.py
Dashboard
- Dashboard UI
frontend/app/page.tsx - API proxy
frontend/app/api/[...path]/route.ts - Dashboard sessions
frontend/lib/dashboard-session.ts - Sensor views
frontend/components/SensorFleet.tsx - Forecast view
frontend/components/ForecastView.tsx - Source and mitigation
frontend/components/SourceIntelligence.tsx
Other components
- Customer application
- Remote sensor
- Operator
Connections 23
- Customer application → Remote sensor: observed in parallel
- Remote sensor → Agent service: observed traffic
- Agent service → Telemetry adapters: collects metadata
- Agent service → Telemetry client: submits batches
- Telemetry client → Central API: sends telemetry
- Central API → Sensor runtime: dispatches telemetry
- Sensor runtime → State aggregation: builds states
- Sensor runtime → State history: maintains history
- Sensor runtime → Forecast inference: requests forecasts
- Forecast inference → Forecast explanation: computes contributions
- Sensor runtime → Source intelligence: prioritizes sources
- Sensor runtime → Mitigation policy: requests recommendations
- Central API → Sensor management: serves sensor views
- Sensor management → Sensor registry: reads sensor records
- Sensor management → Sensor runtime: reads runtime snapshot
- Operator → Dashboard UI: uses
- Dashboard UI → Dashboard sessions: checks session
- Dashboard UI → API proxy: requests data
- API proxy → Central API: forwards requests
- Central API → API proxy: returns responses
- Dashboard UI → Sensor views: displays fleet
- Dashboard UI → Forecast view: displays forecast
- Dashboard UI → Source and mitigation: displays intelligence