Load Balancing Test Results as Management to Prevent Server Down

Abstrak

Increased traffic on a website can lead to server overload, which in turn results in system downtime, slow response, and potential service disruptions that negatively impact user experience. This issue becomes particularly critical for systems that depend on a single main server without an efficient workload distribution mechanism. To address this challenge, this study aims to implement load balancing techniques using the round robin method to distribute incoming requests evenly across multiple backend servers. The research methodology involves system architecture design utilising Docker for containerisation and Nginx as a reverse proxy to manage traffic flow, combined with performance testing conducted using the K6 tool under various simulated traffic scenarios. Experimental results demonstrate that the round robin load balancing method successfully balances the distribution of user requests, reduces the risk of server overload, improves system reliability, and enhances response time performance. Therefore, the proposed approach proves effective in maintaining service availability and optimising overall system performance, especially under high and fluctuating traffic conditions.