The best indicator of the performance of a web application is the average response time. This metric measures how long it takes for the web server to process and deliver a request from the client. It reflects the user's perception of how fast or slow the web application is, and it affects the user's satisfaction, engagement, and conversion. A low average response time means that the web application is responsive and efficient, while a high average response time means that the web application is sluggish and unreliable.
HTTP server error rate, server thread count, and server uptime are not as good indicators of the performance of a web application as the average response time. HTTP server error rate measures how often the web server fails to handle a request and returns an error code, such as 404 (Not Found) or 500 (Internal Server Error).
This metric indicates the reliability and availability of the web application, but it does not capture how fast or slow the web application is. Server thread count measures how many concurrent requests the web server can handle at a given time. This metric indicates the scalability and capacity of the web application, but it does not capture how long each request takes to process. Server uptime measures how long the web server has been running without interruption. This metric indicates the stability and resilience of the web application, but it does not capture how well the web application performs during that time.
References:
10 Key Application Performance Metrics & How to Measure Them - Stackify1 Measuring performance - Learn web development | MDN2 Understanding the Basics of Web Performance | BrowserStack3
14 Important Website Performance Metrics You Should Be Analyzing4
Top 8 Web Application Performance Metrics | MetricFire Blog5
Web Performance Monitoring: A How to Guide for Developers - Stackify6