<![CDATA[This study investigates an integrated public complaint service that combines a non-AI, rule-based WhatsApp chatbot, a web-based administrative dashboard, and a RESTful API to improve early response, status traceability, and ticket-based two-way communication. The system was developed using an Agile Scrum approach, implementing the chatbot in Node.js, the backend services and dashboard in Laravel, and PostgreSQL as the centralized database, while real-time dashboard updates were delivered via WebSocket. Evaluation was conducted through User Acceptance Testing (UAT) for core functional flows and RESTful API load testing using Apache JMeter under gradual-load conditions (Typical Busy, Peak, Stress) and an extreme surge condition (Spike/Burst). The UAT results indicate that all core scenarios passed, covering ticket-based complaint submission, duplicate prevention via a one active ticket per WhatsApp number rule, administrator validation and routing, and real-time conversation synchronization within the ticket context. Under gradual-load conditions, all evaluated endpoints maintained a 0% error rate with sub-second average latency in the range of a few hundred milliseconds, indicating stable baseline behavior as workload increased progressively. Under Spike/Burst, the system remained error-free but latency increased, with average response times of 6,593 ms for create complaint, 18,010 ms for status tracking, 18,321 ms for chat message, and 14,308 ms for mixed load, with throughputs of 7.06 req/s, 2.62 req/s, 2.05 req/s, and 5.90 req/s, respectively. Overall, the results demonstrate end-to-end functional feasibility, stable baseline performance under gradual load, and a resilience boundary under extreme surge, motivating targeted optimization of synchronous processing, history retrieval, and payload serialization to improve Spike/Burst time responsiveness.]]>
Copyrights © 2026
