Performance testing on financial software is massively complex and challenging.
This is the story of how we started to use Splunk (Big Data visualization tool) to help test and debug performance problems at the code level at Murex. During this presentation the old system will be discussed together with its draw backs and how Splunk replaced and enhanced our workflows. My motivation for creating this system was to reduce my workload of analyzing “Big Data” for testing.
Creating a new system and rolling it out to testers and developers can be broken into two main aspects Technical and Human.
The Technical aspect: We had millions and millions of timings logs, displaying them quickly so they are usable for testers and developers. Up to this point, we were using a basic PDF report with graphs with static mathematics. However, when we were able to dynamically graph this data with Splunk the analysis became much quicker and easier. The turnaround time decreased for resolution for testing and development.
The Human aspect: Teams not knowing if a test was running, where the results were and how to access them quickly. This caused a single point of failure on the tester for information. We fixed this issues by developing an open system. We also developed multiple functions in the dashboard to enhance the usability. A user could attach a new environment to Splunk in 20 seconds for live monitoring, they could save a test as a URL to share with colleagues via email, wiki etc..
A success stories is one of the world’s largest “Clearing House”. They process one “Trillion Dollars Daily” with Murex. This is the story of how testers/developers worked together with Splunk to deliver this critical binary in a few months.