Implementasi Autoscaling Untuk Meningkatkan Kinerja Web Server Menggunakan Kubernetes

Sari

The web server is in charge of serving user requests and responding to users every time. The web server requires additional services without interrupting the running processes so that the server continues to function properly. This study aims to improve the performance of the web server in serving user requests to make it more optimal. Serving user requests more optimally can be done with scalability. Scalability is the addition of services on the server but does not affect server performance, this can be done by implementing Autoscaling. Autoscaling is used to improve the optimization of resource usage on a web server in order to maintain stable performance. To implement Autoscaling, a platform called Kubernetes can be used. The results of this study include five parameters, namely Scale Up, Scale Down, Throughput, Response Time, and CPU Usage and determine the durability of the web server. The final result shows that the web server is able to accept request loads of up to 1000000 connections with 10000 connections/second where the virtual web server that has used Autoscaling is able to save resources better than the web server that does not use Autoscaling with the difference from the Throughput parameter is 23.72 Kbps, the difference from the Response Time parameter is 5.16 requests/second, and the difference from the CPU Usage parameter is 5.72% where the average value of the Scale Up Parameter is 90 seconds then for the Scale Down parameter is 280 seconds with the average number of pods added is 5.33.

Keywords - Autoscaling, Webserver, Kubernetes