Performance Evaluation of Docker Containers for Disaster Management Dashboard Web Application

A disaster management dashboard web application is an application used to display streaming data from IoT devices placed in disaster areas. Since its data sent by IoT devices’ Search and Rescue (SAR) agency from a separate place is confidential, it will be stored and processed on a private cloud system. As such, the dashboard web application using Docker containers was built as a solution for data storage and processing in the private cloud server. The ability of the web application to handle and process the user requests, and the performance of serving is the sole responsibility of the container is the key focus of this research. To evaluate its performance, we performed the performance testing of this cloud-based web application using JMeter. The evaluation includes stress testing, endurance testing, and spike testing. These evaluations were executed to monitor the CPU usage, memory, and data transfer for each container of web server, database, messaging, and database administration. The experimental results showed that NGINX applications as web servers used an average of 3% of total CPU usage, while other applications in 3 types of observability scenarios used less than 2%. On the other hand, MySQL as a database application uses an average of 20% of total memory usage, while other applications only use less than 2%. In data transfer measurement, the throughput size tends to be increased linearly according to the number of user requests increases as well. Given the result, this research contributed to providing recommendations for the further development of SAR Agency’s web applications.