[release-4.21] OCPBUGS-134608: Fix pathological events

pkg/monitortestlibrary/pathologicaleventlibrary/duplicated_event_patterns.go

@@ -1035,9 +1035,13 @@ func newVsphereConfigurationTestsRollOutTooOftenEventMatcher(finalIntervals moni
 		return eventInterval.Source == monitorapi.SourceE2ETest &&
 			strings.Contains(eventInterval.Locator.Keys[monitorapi.LocatorE2ETestKey], "snapshot options in clusterCSIDriver")
 	})
+	// The DaemonSet rollout that a test triggers (including the configuration restore at
+	// test teardown) proceeds pod by pod across nodes and can continue for ~15 minutes
+	// after the test itself has finished. Pad the test intervals enough to cover that
+	// rollout tail, so events from it are still attributed to the test.
 	for i := range configurationTestIntervals {
-		configurationTestIntervals[i].To = configurationTestIntervals[i].To.Add(time.Minute * 10)
-		configurationTestIntervals[i].From = configurationTestIntervals[i].From.Add(time.Minute * -10)
+		configurationTestIntervals[i].To = configurationTestIntervals[i].To.Add(time.Minute * 30)
+		configurationTestIntervals[i].From = configurationTestIntervals[i].From.Add(time.Minute * -30)
 	}
 
 	return &OverlapOtherIntervalsPathologicalEventMatcher{
@@ -1046,9 +1050,9 @@ func newVsphereConfigurationTestsRollOutTooOftenEventMatcher(finalIntervals moni
 			locatorKeyRegexes: map[monitorapi.LocatorKey]*regexp.Regexp{
 				monitorapi.LocatorNamespaceKey: regexp.MustCompile(`^openshift-cluster-csi-drivers$`),
 			},
-			messageReasonRegex: regexp.MustCompile(`(.*Create.*|.*Delete.*|.*Update.*)`),
-			messageHumanRegex:  regexp.MustCompile(`(.*Create.*|.*Delete.*|.*Update.*)`),
-			jira:               "https://issues.redhat.com/browse/OCPBUGS-42610",
+			messageReasonRegex: regexp.MustCompile(`(.*Create.*|.*Delete.*|.*Update.*|^ScalingReplicaSet$)`),
+			messageHumanRegex:  regexp.MustCompile(`vmware-vsphere-csi-driver`),
+			jira:               "https://redhat.atlassian.net/browse/OCPBUGS-94112",
 		},
 		allowIfWithinIntervals: configurationTestIntervals,
 	}

pkg/monitortestlibrary/pathologicaleventlibrary/duplicated_events_special_test.go

@@ -2,11 +2,111 @@ package pathologicaleventlibrary
 
 import (
 	"testing"
+	"time"
 
+	v1 "github.com/openshift/api/config/v1"
 	"github.com/openshift/origin/pkg/monitor/monitorapi"
 	"github.com/stretchr/testify/assert"
 )
 
+func TestVsphereConfigurationMatcherUsesLastOccurrence(t *testing.T) {
+	testStart := time.Date(2026, 6, 28, 22, 30, 0, 0, time.UTC)
+	testEnd := testStart.Add(30 * time.Minute)
+	firstTimestamp := testStart.Add(-4 * time.Hour)
+	lastTimestamp := testStart.Add(15 * time.Minute)
+
+	testInterval := monitorapi.NewInterval(monitorapi.SourceE2ETest, monitorapi.Info).
+		Locator(monitorapi.NewLocator().E2ETest("[sig-storage] snapshot options in clusterCSIDriver should restore configuration")).
+		Message(monitorapi.NewMessage().HumanMessage("test interval")).
+		Build(testStart, testEnd)
+
+	tests := []struct {
+		name    string
+		locator monitorapi.Locator
+		reason  monitorapi.IntervalReason
+		message string
+	}{
+		{
+			name:    "daemonset update",
+			locator: monitorapi.NewLocator().DeploymentFromName("openshift-cluster-csi-drivers", "vmware-vsphere-csi-driver-operator"),
+			reason:  "DaemonSetUpdated",
+			message: "Updated DaemonSet.apps/vmware-vsphere-csi-driver-node because it changed",
+		},
+		{
+			name:    "daemonset pod creation",
+			locator: monitorapi.NewLocator().DaemonSetFromName("openshift-cluster-csi-drivers", "vmware-vsphere-csi-driver-node"),
+			reason:  "SuccessfulCreate",
+			message: "Created pod: vmware-vsphere-csi-driver-node-abcde",
+		},
+		{
+			name:    "configuration secret update",
+			locator: monitorapi.NewLocator().DeploymentFromName("openshift-cluster-csi-drivers", "vmware-vsphere-csi-driver-operator"),
+			reason:  "SecretUpdated",
+			message: "Updated Secret/vmware-vsphere-csi-driver-config",
+		},
+		{
+			name:    "deployment scaling",
+			locator: monitorapi.NewLocator().DeploymentFromName("openshift-cluster-csi-drivers", "vmware-vsphere-csi-driver-controller"),
+			reason:  "ScalingReplicaSet",
+			message: "Scaled down replica set vmware-vsphere-csi-driver-controller-abcde from 2 to 1",
+		},
+	}
+
+	matcher := newVsphereConfigurationTestsRollOutTooOftenEventMatcher(monitorapi.Intervals{testInterval})
+	for _, test := range tests {
+		t.Run(test.name, func(t *testing.T) {
+			event := monitorapi.NewInterval(monitorapi.SourceKubeEvent, monitorapi.Info).
+				Locator(test.locator).
+				Message(monitorapi.NewMessage().
+					Reason(test.reason).
+					HumanMessage(test.message).
+					WithAnnotation(monitorapi.AnnotationCount, "25").
+					WithAnnotation("firstTimestamp", firstTimestamp.Format(time.RFC3339)).
+					WithAnnotation("lastTimestamp", lastTimestamp.Format(time.RFC3339))).
+				Build(lastTimestamp, lastTimestamp.Add(time.Second))
+
+			assert.True(t, matcher.Allows(event, v1.HighlyAvailableTopologyMode),
+				"expected the rollout event's last occurrence to be attributed to the configuration test")
+		})
+	}
+
+	// The daemonset rollout triggered by the test (or by the configuration restore at test
+	// teardown) can continue past the end of the test itself; events from that tail must
+	// still be attributed to the configuration test.
+	rolloutTailEvent := monitorapi.NewInterval(monitorapi.SourceKubeEvent, monitorapi.Info).
+		Locator(monitorapi.NewLocator().DaemonSetFromName("openshift-cluster-csi-drivers", "vmware-vsphere-csi-driver-node")).
+		Message(monitorapi.NewMessage().
+			Reason("SuccessfulCreate").
+			HumanMessage("Created pod: vmware-vsphere-csi-driver-node-tlp4z").
+			WithAnnotation(monitorapi.AnnotationCount, "43").
+			WithAnnotation("firstTimestamp", firstTimestamp.Format(time.RFC3339)).
+			WithAnnotation("lastTimestamp", lastTimestamp.Format(time.RFC3339))).
+		Build(testEnd.Add(20*time.Minute), testEnd.Add(20*time.Minute+time.Second))
+	assert.True(t, matcher.Allows(rolloutTailEvent, v1.HighlyAvailableTopologyMode),
+		"expected an event from the post-test rollout tail to be attributed to the configuration test")
+
+	outsideTestWindow := monitorapi.NewInterval(monitorapi.SourceKubeEvent, monitorapi.Info).
+		Locator(monitorapi.NewLocator().DeploymentFromName("openshift-cluster-csi-drivers", "vmware-vsphere-csi-driver-controller")).
+		Message(monitorapi.NewMessage().
+			Reason("ScalingReplicaSet").
+			HumanMessage("Scaled down replica set vmware-vsphere-csi-driver-controller-abcde from 2 to 1").
+			WithAnnotation(monitorapi.AnnotationCount, "25").
+			WithAnnotation("firstTimestamp", firstTimestamp.Format(time.RFC3339))).
+		Build(testEnd.Add(45*time.Minute), testEnd.Add(45*time.Minute+time.Second))
+	assert.False(t, matcher.Allows(outsideTestWindow, v1.HighlyAvailableTopologyMode),
+		"expected an event outside the buffered test interval to remain pathological")
+
+	unrelatedScalingEvent := monitorapi.NewInterval(monitorapi.SourceKubeEvent, monitorapi.Info).
+		Locator(monitorapi.NewLocator().DeploymentFromName("openshift-cluster-csi-drivers", "unrelated-controller")).
+		Message(monitorapi.NewMessage().
+			Reason("ScalingReplicaSet").
+			HumanMessage("some unrelated scaling operation").
+			WithAnnotation(monitorapi.AnnotationCount, "25")).
+		Build(lastTimestamp, lastTimestamp.Add(time.Second))
+	assert.False(t, matcher.Allows(unrelatedScalingEvent, v1.HighlyAvailableTopologyMode),
+		"expected an unrelated scaling event not to match the rollout allowance")
+}
+
 func Test_singleEventThresholdCheck_getNamespacedFailuresAndFlakes(t *testing.T) {
 	namespace := "openshift-etcd-operator"
 	samplePod := "etcd-operator-6f9b4d9d4f-4q9q8"