We run Tsung in Kubernetes to load testing several popular Python WSGI servers to understand the performance of Python servers under high load. This repo described the detailed testing environment.
All the tests are done through Jenkins job and you can find the test results in this spreadsheet. If you wanted to add more target projects to the list or to improve the test cases, feel free to send a PR. You'll find the updated test results after we run the Jenkins job periodically.
Currently we test these Python WSGI servers:
We run these test cases
For each test target, we run load testing for 90 seconds. For the first 60 seconds, 5 new virtual users got spawned every second. Each virtual user will connect to the target server and fire http GET request for 10000 times.
Tsung will dump test stats every 10 seconds. We will gather all the dumped stats for analysis and plotting.
We found Bjoern has the highest QPS at around 10,000/s which is quite impressive. You could find the detailed test results at here.