Noprianto, Nobuo Funabiki, Htoo Htoo Sandi Kyaw, Komang Candra Brata, I Nyoman Darma Kotama, Yan Watequlis Syaifudin, Alfiandi Aulia Rahmadani
Currently, we are developing SEMAR server platform for fast deployments of IoT application systems. However, when many requests simultaneously arrive to SEMAR, it can experience significant performance degradations. Kubernetes is an open-source container orchestration platform designed to manage distributed applications with the essential feature called Horizontal Pod Autoscaler (HPA). It enables dynamical resource adjustment based on resource utilization metrics. In this paper, we propose a performance enhancement of SEMAR using Kubernetes by allowing dynamical scaling pods during load surges and accelerating the recovery time of affected pods. To monitor resource metrics, we use JMeter for traffic loads and Prometheus and Grafana for RAM and CPU usage, as well as processing time in handling requests. In addition, pod downtime and recovery time are measured. The experiment results of the proposal confirmed improved performances with reliable recoveries from failures or disruptions. © 2025 IEEE.
Okayama University, Department of Information and Communication Systems, Okayama, Japan; State Polytechnic of Malang, Department of Information Technology, Malang, Indonesia