Analytics Service

The Analytics Service aggregates data from conversations and system events to generate actionable insights and dashboards.

Overview

It provides real-time metrics on conversation volume, response times, and resolution rates, optionally backed by BigQuery for scalable warehousing.

Key Features

• Dashboard Metrics: Real-time stats for tenant dashboards.
• BigQuery Integration: Optional data warehousing.
• Cost Tracking: Monitor LLM and WhatsApp costs.
• Performance Metrics: Response times and handoff rates.

Architecture

Component	Tech Stack	Description
Language	Python 3.11	Data processing
Framework	FastAPI	Async API
Database	PostgreSQL	Metadata
Warehouse	BigQuery	Large-scale analytics

API Reference

Get Dashboard Metrics

Retrieves aggregated metrics for a date range.

• Endpoint: GET /api/v1/metrics/dashboard
• Query Params: start_date, end_date, outlet_id
• Response:

{
  "conversations": [
    {
      "date": "2025-11-01",
      "total_conversations": 25,
      "resolution_rate": 0.72
    }
  ],
  "costs": [
    {
      "total_llm_cost": 0.008,
      "total_whatsapp_cost": 2.50
    }
  ]
}

Get Tenant Summary

High-level performance summary.

• Endpoint: GET /api/v1/metrics/summary
• Response:

{
  "total_conversations": 750,
  "average_response_time_seconds": 28.5,
  "total_cost": 75.50
}

age summary. for tenant owners.

Integration

• Data Source: Consumes events from other services via message queue (RabbitMQ/Kafka).