How pravinraj213/incident-sentinel works
Incident Sentinel is a small AWS serverless security-monitoring application. An S3 upload invokes its Lambda handler, which asks Rekognition to detect image labels, filters predictions against configured threat-confidence thresholds, and publishes an alert to SNS when any qualify; otherwise it logs a clean result. The graph separates the analysis, decision, and notification stages despite their implementation in one file. The invocation and notification delivery are documented in the README; the source excerpt shows the handler’s service calls. No separate storage subsystem is implemented in the repository.
Main entry points: lambda_function.py Lambda Handler
Components 6
Image Analysis
- Lambda Handler
lambda_function.py - Rekognition
Threat Decision
- Threat Filter
lambda_function.py
Alerting
- SNS Topic
- Email/SMS Recipients
Other components
- S3 Upload
Connections 6
- S3 Upload → Lambda Handler: invokes on upload
README.md - Lambda Handler → Rekognition: requests labels
lambda_function.py - Rekognition → Lambda Handler: returns labels
lambda_function.py - Lambda Handler → Threat Filter: filters predictions
lambda_function.py - Lambda Handler → SNS Topic: publishes qualifying alerts
lambda_function.py - SNS Topic → Email/SMS Recipients: delivers notifications
README.md