There are a number of load testing metrics that tells us about the performance of a web application or website. Performance measurement of an application is portrayed by response metrics that shows user perspective, whereas traffic generated against a web application is shown by volume metrics.

Response Metrics

  • Average Response Time

Response time is calculated from the start of every request and every response to that request, which gives the data of a round trip from the browser and the time it takes to get the response. The average response time is obtained by calculating the mean of all request/response cycle. The obtained value shows the speed of the web application being tested, i.e. how well the site is performing from the user’s perspective. Response time can be measured as:

  • Time to First Byte
  • Time to Last Byte
  • Peak Response Time

Peak response time is similar to average response time that measures the round trip of a request/response cycle. However, peak response time tells us about the interval when there was longest cycle. The peak response time can arise due to various anomalies in the web application, for example, it could be a search result that took a very long time because of the HTML document or database queries.

  • Error Rate

An application under load may give some expected errors while processing requests. Mathematically, the error rate is defined as the percentage of problem requests as compared to all requests. These errors are most common when the application reaches a threshold point that exceeds the ability of a web application to handle more requests. It is significant in deciding the efficiency of an application, as it measures “performance failure” in the application.

Volume Measurements

  • Concurrent Users

It is the most usual way to show the load on an application. This metrics show how many virtual users are active at any specific time. It is different from RPS (Request per Second) in a way that here a virtual user would not be generating requests constantly.

  • Requests per Second

It measures the number of requests that are being sent to the target server, including HTML pages, JavaScript, XML documents, and more. Technically, RPS is affected depending on the number of resources being called from the page.

  • Throughput

Throughput gives you the idea of the bandwidth consumed during the test and is measured in units of Kilobytes per Second. It usually varies depending upon the concurrent users on different tests. However, there could be other reasons as well. A high throughput indicates a high rate of data transfer from the site.

These were some of the most common server load testing metrics that are used. Some of the other mobile app load testing metrics include:

  • Response time
  • Number of transactions passed/failed
  • Load size
  • CPU utilization
  • Memory utilization
  • Wait Time (Average Latency)

 

360logica makes use of open source and commercial tools to execute performance testing. This proves to cost effective and high in standards. We give priority to the client’s requirements and select tools following industry standards based on that. Our performance testing services are also provided through client server and web applications. This allows us to perform load and stress testing on Windows, Linux, and other platforms.