+++ This bug was initially created as a clone of Bug #1932619 +++ +++ This bug was initially created as a clone of Bug #1932618 +++ https://github.com/openshift/cluster-baremetal-operator/pull/110 merged containing a failing alert ClusterOperatorBaremetalDown and TargetDown. This was the passing run https://prow.ci.openshift.org/view/gs/origin-ci-test/pr-logs/pull/openshift_cluster-baremetal-operator/110/pull-ci-openshift-cluster-baremetal-operator-master-e2e-agnostic/1364398641615212544 The query in the alert test "shouldn't report any alerts in firing state apart from Watchdog and AlertmanagerReceiversNotConfigured" is subtly wrong. For 4.6, 4.7, and 4.8, we can remove the broken filter clause because both KubeAPILatencyHigh and KubePodCrashLooping on kcm namespace were fixed. In the future we must use "unless X" instead of joining with a "-" because of the way the series match. After this fix we will be correctly enforcing "no alerts may fire during a CI test run".
This has been fixed in OpenShift 4.8. I'm closing this BZ, if a backport to 4.6 is requested please reopen.