How aditishosur/kyros-ai works
PulseOps is an API-operations intelligence platform. Its principal workflow is to inspect telemetry, forecast workload, detect anomalies, rank supported causes, and estimate scenario interventions; the Streamlit demo presents results through a FastAPI integration layer. The README also describes offline forecasting, anomaly, and scenario research. Research components are mapped to their documented directories; their detailed wiring is not established by the sampled source.
Main entry points: main.py FastAPI routes · database.py SQLite telemetry · ml.py Forecast and anomaly · analytics.py Telemetry analytics
Components 18
Demo interface
- Streamlit demo
frontend/app.py - HTTP client
frontend/client.py
API integration
- FastAPI routes
backend/main.py - API schemas
backend/schemas.py
Operational intelligence
- Telemetry analytics
backend/services/analytics.py - Risk scoring
backend/services/risk.py - Forecast and anomaly
backend/services/ml.py - Root-cause ranking
backend/services/root_cause.py - Intervention simulation
backend/services/simulation.py
Telemetry storage
- SQLite telemetry
backend/database.py - Persistence models
backend/models.py - Database initialization
backend/bootstrap.py
Offline research
- Synthetic telemetry
data/generate_data.py - Telemetry ETL demo
extras/etl_pipeline.py - Forecast experiments
experiments/forecasting - Anomaly evaluation
experiments/anomaly - Scenario generation
experiments/scenarios
Other components
- Operations user
Connections 22
- Operations user → Streamlit demo: uses
- Streamlit demo → HTTP client: requests data
- HTTP client → FastAPI routes: HTTP requests
- FastAPI routes → API schemas: validates payloads
- FastAPI routes → SQLite telemetry: opens sessions
- FastAPI routes → Database initialization: initializes on startup
- Database initialization → SQLite telemetry: initializes store
- FastAPI routes → Persistence models: queries and updates
- FastAPI routes → Telemetry analytics: calls analytics
- FastAPI routes → Forecast and anomaly: calls forecast and anomaly
- FastAPI routes → Root-cause ranking: calls diagnosis
- FastAPI routes → Intervention simulation: runs simulation
- Telemetry analytics → SQLite telemetry: reads telemetry
- Risk scoring → Forecast and anomaly: scores forecast risk
- Forecast and anomaly → Telemetry analytics: loads telemetry
- Forecast and anomaly → SQLite telemetry: stores predictions
- Root-cause ranking → SQLite telemetry: reads telemetry
- Intervention simulation → SQLite telemetry: reads telemetry
- Synthetic telemetry → Telemetry ETL demo: generates input
- Telemetry ETL demo → SQLite telemetry: loads telemetry
- Forecast experiments → Anomaly evaluation
- Scenario generation → Anomaly evaluation: generates labeled episodes