Parallelism Failures: Common Denominators



We experience intermittent failures in our test suite, possibly due to effects or interactions with previous specs that have run. CircleCi provides parallelism, and since it knows which specs were run in which containers, might have a history of failure rate for each spec in relation to which other specs it was run with.

It would be great if there were an easy way to analyze this data, so I could check to see if a specific random failure in file XX always happens when running in a container alongside file YY, but does not always fail when in the same container as file ZZ.