Description
We currently set `maxParallelForks` to the number returned by `Runtime.availableProcessors`.
tasks.withType(Test) {
maxParallelForks = Runtime.runtime.availableProcessors()
}
This returns the number of logical cores (including hyperthreaded cores) in the machine.
This is usually OK when running the tests locally, but the Apache Jenkins slaves run 2 to 3 jobs simultaneously causing a higher number of timing related failures.
A potential solution is to allow `maxParallelForks` to be set via a Gradle property and use that property to set it to an appropriate value when the build is run from Jenkins.