Implementasi Monitoring Suhu dan Kelembapan Rak Server Berbasis Raspberry Pi serta Sensor LH-WS46-HW dengan Protokol Modbus RTU

Abstract

Pemantauan suhu dan kelembapan di ruang server PT Tera Tekno Infotama secara manual berisiko memperlambat penanganan gangguan. Penelitian ini membangun sistem monitoring rak server berbasis Raspberry Pi 3B v1.2 dan sensor industri LH-WS46-HW dengan protokol Modbus RTU melalui RS485 secara real-time. Data dibaca setiap 10 detik, disimpan ke basis data PostgreSQL pada VPS, divisualisasikan pada dashboard Grafana, dan terintegrasi dengan bot Telegram untuk notifikasi otomatis kondisi kritis (suhu = 30 °C, kelembapan = 60% RH, atau timeout > 60 detik). Pengujian kesesuaian terhadap sensor pembanding XY-MD02 pada 13.562 pasang data menghasilkan error 2,25% (suhu) dan 3,59% (kelembapan), nilai RMSE 0,63 °C dan 1,93% RH, serta koefisien korelasi 0,9095 dan 0,9642. Selisih pembacaan sensor LH-WS46-HW terhadap sensor pembanding bervariasi antara 0,6 °C hingga 0,8 °C untuk suhu dan 1,3% RH hingga 3,1% RH untuk kelembapan. Seluruh lima skenario uji fungsional berhasil dilakukan. Rata-rata waktu penyimpanan data sebesar 86,47 ms, penampilan data 62,67 ms, dan pengiriman notifikasi 16,3 detik. Sistem ini terbukti efektif dan layak diimplementasikan untuk pemantauan rak server secara real-time.

Manual monitoring of temperature and humidity in the server room of PT Tera Tekno Infotama risks delaying responses to system issues. This study constructs a real-time server rack monitoring system utilizing a Raspberry Pi 3B v1.2 and an industrial-grade LH-WS46-HW sensor via the Modbus RTU protocol over RS485. Data are collected every 10 seconds, stored in a PostgreSQL database on a VPS, visualized on a Grafana dashboard, and integrated with Telegram bot notifications for critical alerts (temperature = 30 °C, humidity = 60% RH, or timeout > 60 seconds). Agreement testing against an XY-MD02 comparison sensor across 13,562 data pairs yielded errors of 2.25% for temperature and 3.59% for humidity, RMSE values of 0.63 °C and 1.93% RH, correlation coefficients of 0.9095 and 0.9642, and mean differences of -0.60 °C and +1.85% RH. Differences from the comparison sensor varied from 0.6 °C to 0.8 °C for temperature and from 1.3% RH to 3.1% RH for humidity. All five functional testing scenarios succeeded. The average data storage time was 86.47 ms, data display latency was 62.67 ms, and notification delivery time was 16.3 seconds. The system is proven effective and feasible for real-time server rack monitoring.