Establish statistically significant and consistent performance tests

[mousetraps]

Currently, there is a lot of variance with performance tests - we need to set up an environment where:

• we can understand the impact of our work on multiple machine configurations
• we can minimize variance between test runs and have higher confidence that a potential performance optimization is actually adding value rather than unnecessarily complicating the codebase
• the tests should run continuously so we can detect issues as soon as possible