I made some changes recently to try and increase test coverage for some code in the concurrency module. Unfortunately I've repeatedly had to back out the commits due to their seemingly resulting in "flaky failures" in CI. I was wondering if anyone had any advice as to strategies to use to try and reproduce flaky test failures locally. I generally assume CI machines are probably at far higher CPU utilization than my personal laptop, so I tried a tool that the internet suggested would be able to synthetically increase CPU load[1]. However I've still not been able to induce the test failures. Has anyone ever had any success with this sort of thing?
stress-ngwas the specific one – not super familiar with it, but it did appear to peg my CPU(s) when enabled. ↩︎