Support Ticket System – Tech Intern Assessment

Overview

This project is a full-stack Support Ticket System built as part of the Tech Intern – Full Stack Development assessment.

The system allows users to:

• Create support tickets
• Automatically classify tickets using an LLM
• Override suggested category and priority
• Filter and search tickets
• Update ticket status
• View aggregated ticket statistics

The entire application runs using Docker with a single command.

---

Tech Stack

Backend

• Django
• Django REST Framework
• PostgreSQL
• OpenRouter (LLM integration)
• Django ORM (aggregate, annotate)

Frontend

• React (Create React App)
• Fetch API
• Component-based architecture

Infrastructure

• Docker
• Docker Compose
• PostgreSQL volume persistence
• Database healthcheck configuration

---

Setup Instructions

1️ Create Environment File

Create a .env file in the root directory:

OPENROUTER_API_KEY=your_api_key_here

(The project uses OpenRouter for LLM integration.)

2️ Run the Application

docker compose up --build

3️ Access the Application

Frontend: http://localhost:3000

Backend API: http://localhost:8000/api/

LLM Integration

The endpoint:

POST /api/tickets/classify/

• Accepts a ticket description as input
• Sends a request to the OpenRouter Chat Completion API
• Uses structured JSON prompting to enforce consistent output format
• Returns suggested:
  • Category (billing, technical, account, general)
  • Priority (low, medium, high, critical)

Environment variable used: OPENROUTER_API_KEY

Graceful fallback is implemented: If the LLM request fails (e.g., missing API key, network issue, invalid response), the system defaults to:

• category: general
• priority: low

This ensures ticket creation is never blocked due to LLM failure.

Stats Endpoint

The endpoint:

GET /api/tickets/stats/

Uses database-level aggregation:

• Count()
• Avg()
• TruncDate()
• annotate()
• aggregate()

No Python-level loops are used for aggregation.

Key Design Decisions

• Docker-first architecture
• PostgreSQL volume persistence
• Healthcheck-based service dependency
• Environment variable-based secret handling
• Graceful LLM error handling
• Separation of concerns (service layer in frontend)
• Clean REST API structure

Features Implemented

• Ticket CRUD API
• Filtering (category, priority, status)
• Search (title + description)
• Live status updates (without page reload)
• Stats dashboard with auto-refresh
• LLM-based auto-classification
• Fully containerized deployment

Submission Notes

• Entire project runs with docker compose up --build
• No manual setup required (except API key)
• .git directory included for commit history review

API Endpoints

Base URL: http://localhost:8000/api/

Available endpoints:

• GET /tickets/ → List all tickets
• POST /tickets/ → Create a ticket
• PATCH /tickets/{id}/ → Update ticket status
• GET /tickets/?category=... → Filter by category
• GET /tickets/?priority=... → Filter by priority
• GET /tickets/?status=... → Filter by status
• GET /tickets/?search=... → Search by title or description
• GET /tickets/stats/ → Get aggregated statistics
• POST /tickets/classify/ → Classify ticket using LLM