How muhil-amuthan/netsentry-ai works
NetSentry-AI is a telecom NOC incident-triage application. Its principal workflow turns alerts into normalized, deduplicated, correlated incidents; scores impact, retrieves runbook evidence, then recommends actions or escalates when evidence is insufficient. The API and dashboard expose incident views and an “Ask NetSentry” interface. The README documents the end-to-end stages; source excerpts and the SOURCE INDEX support the module coverage and selected connections, but not every internal call sequence.
Main entry points: api.py FastAPI routes · database.py In-memory state · runbook_engine.py Runbook retrieval · app.js NOC dashboard
Components 18
NOC interface
- NOC dashboard
frontend/app.js - Ask NetSentry
src/nlp_handler.py
API and state
- FastAPI routes
src/api.py - In-memory state
src/database.py
Incident intelligence
- Domain models
src/models.py - Scenario alerts
src/generator.py - Normalize and dedupe
src/processor.py - Network topology
src/topology.py - Correlation engine
src/scorer.py - Priority engine
src/priority.py
Evidence and response
- Runbook documents
data/runbooks - Runbook retrieval
src/runbook_engine.py - FAISS index
data/faiss_index/index.faiss - Grounded recommendations
src/generator.py - Human escalation
src/escalation.py
Other components
- NOC engineer
- Alert feed
- Gemini AI
Connections 19
- NOC engineer → NOC dashboard: views triage
README.md - NOC dashboard → FastAPI routes: uses API
README.md - Alert feed → FastAPI routes: submits alerts
- FastAPI routes → In-memory state: queries state
src/api.py - In-memory state → Scenario alerts: loads scenarios
README.md - In-memory state → Normalize and dedupe: processes alerts
README.md - In-memory state → Correlation engine: correlates alerts
README.md - Correlation engine → Network topology: checks connectivity
src/scorer.py - In-memory state → Priority engine: scores incidents
README.md - Priority engine → Network topology: uses impact context
src/priority.py - FastAPI routes → Runbook retrieval: analyzes incidents
src/api.py - Runbook retrieval → Runbook documents: loads evidence
src/runbook_engine.py - Runbook retrieval → FAISS index: retrieves matches
README.md - Runbook retrieval → Gemini AI: embeds queries
src/runbook_engine.py - FastAPI routes → Grounded recommendations: requests analysis
src/api.py - FastAPI routes → Human escalation: evaluates escalation
src/api.py - FastAPI routes → Ask NetSentry: answers questions
src/api.py - Ask NetSentry → In-memory state: queries incident state
src/nlp_handler.py - In-memory state → NOC dashboard: serves incident views
README.md