Tracking performance issues during load testing

One of the more challenging aspects of load testing is isolating the root cause of problems. You generate load on your system and 20 minutes into the test you identify a spike in response time. What’s next? To identify the cause of the problem, you want to track the corresponding server-side events at that point in time.

The Dynatrace integration implemented in WebLOAD, our enterprise load testing tool, lets you do just that. You execute a test scenario that generates load on your system. While analyzing one of WebLOAD’s reports and identifying a performance problem, you can switch to Dynatrace using a single click. You are automatically placed at the exact context of your test, which lets you to conduct a deep analysis of the relevant events.

WebLOAD lets you perform this deep analysis starting from a specific time frame in your load test, or by clicking an error message received from the server. Below is a use case that demonstrates one such scenario.

The WebLOAD graph below clearly highlights an issue with the login transaction measured in the test. While some logins are normal, others show a spike in the response time. Right-clicking the spike on the WebLOAD graph lets you switch to Dynatrace to identify the cause of the problem.

Within Dynatrace, you are automatically placed within the context of the ‘Login’ transaction. Viewing all the Pure Paths within the time window, you can identify the variances in the response time for different login transactions. The slowest ones shown in the screen capture below are for the user called ‘Monica’.

Using the ‘Response Time Hotspot’ you can drill down further to determine that the lengthy login is primarily due to the slow database access.

Switching to the ‘Database’ view you can notice that 927 transactions are sent per login request (!) revealing a system problem.

This integration, which enables tracing performance issues all the way down to an invoked method, cuts root-cause identification time significantly, helping accelerate processes with agile teams and continuous delivery pipelines.

About the Author

David Buch is Director of Innovation at RadView. David has 25 years’ experience in the software industry with more than 15 years in the performance domain, managing R&D and QA groups and assisting customers to achieve their goals.