When analyzing performance test results it’s important to visualize critical metrics to find potential performance issues. With a simple formula, the throughput of any transaction can be calculated based on a few variables.
Transaction Throughput = Number of Users / (Response Time + Think-Time)
The Apica Load Test Portal and the result details page has been improved in version 5.5 so the user can access response times for all pages and transactions in a test, in relation to number of concurrent users. This new feature makes it easier to analyze the result and find maximum throughput.
Two new graphs on the Result Detail page
Detailed Result Graph
The detailed result graph above shows a standard test application where the number of users is ramped-up from 1 to 50 users (green trend). The test application has 6 pages and one of the pages named “Page #6 Throughput” (pink trend) has a significant increase in response time to about 30 users. Response time is almost flat around 2.5 seconds, but with more than 30 concurrent users it increases up to 7.5 seconds.
The conclusion from this specific test example shows the system can handle a maximum of 30 users without any increase in response time.
Scaling and Zooming, Min, Avg. and Max Response Time Trends
The scaling of the page and transaction response time trends starts at zero as default.
The user can then zoom in both the time (X axle) and the scaling of the response time (Y axle) to highlight a specific timeframe. By scaling both axles it is now much easier to highlight a specific part of the trends.
We have also added minimum and maximum response times for each page to identify spikes. When the user zooms in on a specific timeframe avg., min and max response time is shown for that specific time frame.
This can be helpful when answering questions like “What was the response time with a specific number of users during ramp-up of the load.”