While waiting for the developer to fix the root cause, the team identified areas for potential improvement. The team quickly identified excessive SQL queries were to blame for the application’s poor performance, and reported those results to the application developer. Each test scenario had up to 2,400 page loads - with an average of 600 database calls on each page load.īlazeMeter allowed them to rapidly see that the app was not successfully scaling, and data from New Relic enabled them to pinpoint why. Over a two-week period, they ran 150 tests for up to 100 concurrent users. Using JMeter, BlazeMeter, and the application monitoring tool New Relic, the team built a test plan, executed the tests and analyzed the results. At this point the team turned to BlazeMeter. The team found that although Selenium is good for functional testing, it’s problematic when it comes to load testing. However, they also needed to see sessions on a fully functional browser - or as close to one as possible. As some members of the 23-strong team had experience with JMeter, they started with this open source tool, believing it to be the fastest way to build a reasonable test case. “Serving the page isn’t good enough, you have to serve it fast enough for conversion.” - Upic Chief Technology Officer David Burley.īefore the start of the 2016 campaign season, the Upic team, led by David Burley, decided to look for a tool to test and analyze why the donation system was so slow.
Slow pages meant that donors couldn’t pledge. During the campaign season in 2015, the performance of this key app was poor.
0 kommentar(er)
