According to https://ci.wildfly.org/project.html?projectId=WF_PullRequest&buildTypeId=&tab=testDetails&testNameId=-7723124563322321950&order=TEST_STATUS_DESC&branch_WF_PullRequest=__all_branches__&itemsCount=50 in the last month there have been 9 failures of ManagedExecutorServiceMetricsTestCase.testManagedExecutorServiceManagement with this failure:

java.lang.AssertionError
	at deployment.arquillian-service//org.junit.Assert.fail(Assert.java:86)
	at deployment.arquillian-service//org.junit.Assert.assertTrue(Assert.java:41)
	at deployment.arquillian-service//org.junit.Assert.assertTrue(Assert.java:52)
	at deployment.ManagedExecutorServiceMetricsTestCase.jar//org.jboss.as.test.integration.ee.concurrent.ManagedExecutorServiceMetricsTestCase.testRuntimeStats(ManagedExecutorServiceMetricsTestCase.java:238)
	at deployment.ManagedExecutorServiceMetricsTestCase.jar//org.jboss.as.test.integration.ee.concurrent.ManagedExecutorServiceMetricsTestCase.testManagedExecutorServiceManagement(ManagedExecutorServiceMetricsTestCase.java:108)

First failure was reported against https://github.com/wildfly/wildfly/pull/13414 but that was a docs only change so not relevant. It occurred on July 17 but there weren't any merge commits to master in the period between Jul 12 and Jul 20 so any change in master causing this isn't obvious.

The failure is because the test checks that the # of active threads is 0 immediately after it does some activity, so it may just be a timing issue that's always been possible and just started happening recently for env reason on ci.wildfly.org.

A general comment is this test would be better if its asserts reported more info about what it read from the server. Use assertEquals and report the expected and wrong value.