Overview & Drilldown

Project Overview & Session Drilldown

The AI-Horizon console provides dashboards to aggregate multi-project analytics and drill down into specific agent executions. These dashboards are powered by the Project Overview (ExistingProject.jsx) and the Session Drilldown (Drilldown.jsx) modules.


1. Project Overview Dashboard

The Project Overview dashboard (available in the dashboard at /existingproject) serves as the administrative command center. It monitors high-level operational statistics across all workspace project environments.

Key Visualizations & Features

  • Total Token Consumption: Accumulates and displays count charts showing prompt vs. completion tokens.
  • Cost Metrics: Sums agent invocation costs to display aggregated billing data.
  • Success Rate Metrics: Highlights agent status distribution (successes vs. failures) using Recharts Pie cells.
  • Project Filter Panels: Filters transaction histories by dates (From/To) and searches across specific agents.

2. Session Drilldown Telemetry

The Session Drilldown panel (available in the dashboard at /drilldown) displays telemetry logs for a selected single run transaction.

Features

  • Interactive Trace Timelines: Chronologically renders nested parent-child events.
  • Detailed Parameter Inspection: Allows developers to view input files, request payloads, model prompts, and agent system instructions.
  • Parsed Completion Output: Displays final responses and structural JSON outputs (e.g. medical summaries or fitness plans).
  • Citations Auditor: Lists the specific document source chunks fetched by the vector store during RAG processing.

3. Database & REST API Reference

The analytics dashboards fetch and update data using the following endpoints:

1. Add/Update Drilldown Record (POST /api/drilldown)

Creates or updates a session transaction log when an agent completes a run.

  • Request Payload:

    • session_id: Unique session identifier.
    • agent_type: Agent category (e.g., PDFChatBot, Voicebot).
    • aih_api_key & openai_api_key: Keys used for the run.
    • vector_store_params: Vector database parameters.
    • voiceBot_response: Response object from the bot run.
    • file_path: Path of the reference document.
  • Response:

    { "msg": "Data updated successfully" } // or "File uploaded successfully"

2. Fetch Drilldown Records (GET /api/drilldownFetchData)

Retrieves all historical execution logs for project analysis.

  • Response:
    [
      {
        "session_id": "session-uuid-001",
        "agent_type": "PDFChatBot",
        "voiceBot_response": {
          "Session_Details": {
            "overall_cost": 0.015,
            "overall_time_elapsed": "0h 0m 1.5s"
          }
        }
      }
    ]

3. Delete Session Log (DELETE /api/deleteDrillDownData)

Deletes an execution record from the history table.

  • Query Parameters:
    • sessionID: Session ID to delete.
  • Response:
    { "msg": "Record deleted successfully" }