When a task has passed/failed enough times to be considered flaky, quit running it.

For instance, if a test failed 10 times, passed 10 times. Even if the test was run the full 100 iterations, it should still be flaky.

In this case the data point should be good enough to continue.