Real-Time Healthcare Notifications & AI Agent Integration

Critical healthcare environments require instant notifications. At RS Pertamina Dumai (RSPD), delays in dispatching task queues or pharmacy notifications directly affect medical services. The goal was to build a robust, zero-latency notification infrastructure.

The Solution

1. MQTT Telemetry Integration: Configured an MQTT message broker to capture database events and broadcast updates securely.
2. AI-Driven Telegram Interface: Integrated a custom LLM-powered assistant (Gemini/OpenClaw AI) to let administrators query metrics, patient queues, and service logs in natural language.
3. Daemon Service Stability: Deployed the solution as a systemd service (gemini-bot) with ALSA driver fault-handling, ensuring 24/7 background persistence and automated recovery.

Technical Architecture

• Language & Runtime: Python daemon running on Linux.
• Protocol & Queues: MQTT broker for lightweight, low-overhead messaging.
• Alert Channels: Telegram Bot API with granular user/role access verification.
• Database Engine: MySQL/SQLite change tracking scripts (permintaan_semua.py).

Key Outcomes

• Immediate Response: Reduced notification delay from 15 minutes to under 2 seconds.
• Enhanced Visibility: Real-time updates delivered instantly to doctor and coordinator group chats.
• High Availability: Automated service daemon restart hooks, achieving 99.9% uptime.