e2e test issues solved (#6271)

* Improve e2e tests Signed-off-by: Jorge Turrado <[email protected]> * . Signed-off-by: Jorge Turrado <[email protected]> * . Signed-off-by: Jorge Turrado <[email protected]> * . Signed-off-by: Jorge Turrado <[email protected]> * increase the elastic messages Signed-off-by: Jorge Turrado <[email protected]> * fix namespace duplications Signed-off-by: Jorge Turrado <[email protected]> --------- Signed-off-by: Jorge Turrado <[email protected]> Signed-off-by: Jorge Turrado <[email protected]>
kedacore · Nov 1, 2024 · 4376101 · 4376101
1 parent b2ce95d
commit 4376101
Show file tree

Hide file tree

Showing 29 changed files with 109 additions and 77 deletions.
diff --git a/.github/workflows/pr-e2e.yml b/.github/workflows/pr-e2e.yml
@@ -181,7 +181,7 @@ jobs:
       - name: Scale cluster
         run: make scale-node-pool
         env:
-          NODE_POOL_SIZE: 2
+          NODE_POOL_SIZE: 3
           TEST_CLUSTER_NAME: keda-e2e-cluster-pr
 
       - name: Run end to end tests

diff --git a/.github/workflows/template-main-e2e-test.yml b/.github/workflows/template-main-e2e-test.yml
@@ -26,7 +26,7 @@ jobs:
       - name: Scale cluster
         run: make scale-node-pool
         env:
-          NODE_POOL_SIZE: 2
+          NODE_POOL_SIZE: 3
 
       - name: Run end to end tests
         env:

diff --git a/pkg/scalers/kubernetes_workload_scaler.go b/pkg/scalers/kubernetes_workload_scaler.go
@@ -33,7 +33,7 @@ var phasesCountedAsTerminated = []corev1.PodPhase{
 
 type kubernetesWorkloadMetadata struct {
 	PodSelector     string  `keda:"name=podSelector,     order=triggerMetadata"`
-	Value           float64 `keda:"name=value,           order=triggerMetadata"`
+	Value           float64 `keda:"name=value,           order=triggerMetadata, default=0"`
 	ActivationValue float64 `keda:"name=activationValue, order=triggerMetadata, default=0"`
 
 	namespace      string
@@ -72,17 +72,13 @@ func NewKubernetesWorkloadScaler(kubeClient client.Client, config *scalersconfig
 
 func parseKubernetesWorkloadMetadata(config *scalersconfig.ScalerConfig) (kubernetesWorkloadMetadata, error) {
 	meta := kubernetesWorkloadMetadata{}
-	err := config.TypedConfig(&meta)
-	if err != nil {
-		return meta, fmt.Errorf("error parsing kubernetes workload metadata: %w", err)
-	}
-
 	meta.namespace = config.ScalableObjectNamespace
 	meta.triggerIndex = config.TriggerIndex
 	meta.asMetricSource = config.AsMetricSource
 
-	if meta.asMetricSource {
-		meta.Value = 0
+	err := config.TypedConfig(&meta)
+	if err != nil {
+		return meta, fmt.Errorf("error parsing kubernetes workload metadata: %w", err)
 	}
 
 	selector, err := labels.Parse(meta.PodSelector)

diff --git a/tests/helper/helper.go b/tests/helper/helper.go
@@ -253,6 +253,7 @@ func DeleteNamespace(t *testing.T, nsName string) {
 		err = nil
 	}
 	assert.NoErrorf(t, err, "cannot delete kubernetes namespace - %s", err)
+	DeletePodsInNamespace(t, nsName)
 }
 
 func WaitForJobSuccess(t *testing.T, kc *kubernetes.Clientset, jobName, namespace string, iterations, interval int) bool {
@@ -744,6 +745,17 @@ func DeletePodsInNamespaceBySelector(t *testing.T, kc *kubernetes.Clientset, sel
 	assert.NoErrorf(t, err, "cannot delete pods - %s", err)
 }
 
+// Delete all pods in namespace
+func DeletePodsInNamespace(t *testing.T, namespace string) {
+	err := GetKubernetesClient(t).CoreV1().Pods(namespace).DeleteCollection(context.Background(), metav1.DeleteOptions{
+		GracePeriodSeconds: ptr.To(int64(0)),
+	}, metav1.ListOptions{})
+	if errors.IsNotFound(err) {
+		err = nil
+	}
+	assert.NoErrorf(t, err, "cannot delete pods - %s", err)
+}
+
 // Wait for Pods identified by selector to complete termination
 func WaitForPodsTerminated(t *testing.T, kc *kubernetes.Clientset, selector, namespace string, iterations, intervalSeconds int) bool {
 	for i := 0; i < iterations; i++ {

diff --git a/tests/internals/cache_metrics/cache_metrics_test.go b/tests/internals/cache_metrics/cache_metrics_test.go
@@ -160,8 +160,8 @@ func testCacheMetricsOnPollingInterval(t *testing.T, kc *kubernetes.Clientset, d
 
 	// Metric Value = 8, DesiredAverageMetricValue = 2
 	// should scale in to 8/2 = 4 replicas, irrespective of current replicas
-	assert.True(t, WaitForDeploymentReplicaReadyCount(t, kc, deploymentName, testNamespace, 4, 60, 1),
-		"replica count should be 4 after 1 minute")
+	assert.True(t, WaitForDeploymentReplicaReadyCount(t, kc, deploymentName, testNamespace, 4, 60, 3),
+		"replica count should be 4 after 3 minute")
 
 	// Changing Metric Value to 4, but because we have a long polling interval, the replicas number should remain the same
 	data.MonitoredDeploymentReplicas = 4
@@ -196,8 +196,8 @@ func testDirectQuery(t *testing.T, kc *kubernetes.Clientset, data templateData)
 
 	// Metric Value = 8, DesiredAverageMetricValue = 2
 	// should scale in to 8/2 = 4 replicas, irrespective of current replicas
-	assert.True(t, WaitForDeploymentReplicaReadyCount(t, kc, deploymentName, testNamespace, 4, 60, 1),
-		"replica count should be 4 after 1 minute")
+	assert.True(t, WaitForDeploymentReplicaReadyCount(t, kc, deploymentName, testNamespace, 4, 60, 3),
+		"replica count should be 4 after 3 minute")
 
 	// Changing Metric Value to 4, deployment should scale to 2
 	data.MonitoredDeploymentReplicas = 4

diff --git a/tests/internals/eventemitter/azureeventgridtopic/azureeventgridtopic_test.go b/tests/internals/eventemitter/azureeventgridtopic/azureeventgridtopic_test.go
@@ -265,21 +265,25 @@ func checkMessage(t *testing.T, count int, client *azservicebus.Client) {
 	if err != nil {
 		assert.NoErrorf(t, err, "cannot create receiver - %s", err)
 	}
-	defer receiver.Close(context.TODO())
-
-	messages, err := receiver.ReceiveMessages(context.TODO(), count, nil)
-	assert.NoErrorf(t, err, "cannot receive messages - %s", err)
-	assert.NotEmpty(t, messages)
+	defer receiver.Close(context.Background())
 
+	// We try to read the messages 3 times with a second of delay
+	tries := 3
 	found := false
-	for _, message := range messages {
-		event := messaging.CloudEvent{}
-		err = json.Unmarshal(message.Body, &event)
-		assert.NoErrorf(t, err, "cannot retrieve message - %s", err)
-		if expectedSubject == *event.Subject &&
-			expectedSource == event.Source &&
-			expectedType == event.Type {
-			found = true
+	for i := 0; i < tries && !found; i++ {
+		messages, err := receiver.ReceiveMessages(context.Background(), count, nil)
+		assert.NoErrorf(t, err, "cannot receive messages - %s", err)
+		assert.NotEmpty(t, messages)
+
+		for _, message := range messages {
+			event := messaging.CloudEvent{}
+			err = json.Unmarshal(message.Body, &event)
+			assert.NoErrorf(t, err, "cannot retrieve message - %s", err)
+			if expectedSubject == *event.Subject &&
+				expectedSource == event.Source &&
+				expectedType == event.Type {
+				found = true
+			}
 		}
 	}
 

diff --git a/tests/internals/idle_replicas/idle_replicas_test.go b/tests/internals/idle_replicas/idle_replicas_test.go
@@ -147,8 +147,8 @@ func testScaleOut(t *testing.T, kc *kubernetes.Clientset) {
 
 	t.Log("--- scale to max replicas ---")
 	KubernetesScaleDeployment(t, kc, monitoredDeploymentName, 4, testNamespace)
-	assert.True(t, WaitForDeploymentReplicaReadyCount(t, kc, deploymentName, testNamespace, 4, 60, 1),
-		"replica count should be 4 after 1 minute")
+	assert.True(t, WaitForDeploymentReplicaReadyCount(t, kc, deploymentName, testNamespace, 4, 60, 3),
+		"replica count should be 4 after 3 minute")
 }
 
 func testScaleIn(t *testing.T, kc *kubernetes.Clientset) {

diff --git a/tests/internals/restore_original/restore_original_test.go b/tests/internals/restore_original/restore_original_test.go
@@ -138,8 +138,8 @@ func testScale(t *testing.T, kc *kubernetes.Clientset, data templateData) {
 	t.Log("--- testing scaling ---")
 	KubectlApplyWithTemplate(t, data, "scaledObjectTemplate", scaledObjectTemplate)
 
-	assert.True(t, WaitForDeploymentReplicaReadyCount(t, kc, deploymentName, testNamespace, 4, 60, 1),
-		"replica count should be 4 after 1 minute")
+	assert.True(t, WaitForDeploymentReplicaReadyCount(t, kc, deploymentName, testNamespace, 4, 60, 3),
+		"replica count should be 4 after 3 minute")
 }
 
 func testRestore(t *testing.T, kc *kubernetes.Clientset, data templateData) {

diff --git a/tests/internals/scaled_job_validation/scaled_job_validation_test.go b/tests/internals/scaled_job_validation/scaled_job_validation_test.go
@@ -13,7 +13,7 @@ import (
 )
 
 const (
-	testName = "scaled-object-validation-test"
+	testName = "scaled-job-validation-test"
 )
 
 var (

diff --git a/tests/internals/scaling_strategies/eager_scaling_strategy_test.go b/tests/internals/scaling_strategies/eager_scaling_strategy_test.go
@@ -100,8 +100,11 @@ func TestScalingStrategy(t *testing.T) {
 	})
 
 	RMQInstall(t, kc, rmqNamespace, user, password, vhost, WithoutOAuth())
-	CreateKubernetesResources(t, kc, testNamespace, data, templates)
+	// Publish 0 messges but create the queue
+	RMQPublishMessages(t, rmqNamespace, connectionString, queueName, 0)
+	WaitForAllJobsSuccess(t, kc, rmqNamespace, 60, 1)
 
+	CreateKubernetesResources(t, kc, testNamespace, data, templates)
 	testEagerScaling(t, kc)
 }
 
@@ -121,14 +124,17 @@ func getTemplateData() (templateData, []Template) {
 func testEagerScaling(t *testing.T, kc *kubernetes.Clientset) {
 	iterationCount := 20
 	RMQPublishMessages(t, rmqNamespace, connectionString, queueName, 4)
+	WaitForAllJobsSuccess(t, kc, rmqNamespace, 60, 1)
 	assert.True(t, WaitForScaledJobCount(t, kc, scaledJobName, testNamespace, 4, iterationCount, 1),
 		"job count should be %d after %d iterations", 4, iterationCount)
 
 	RMQPublishMessages(t, rmqNamespace, connectionString, queueName, 4)
+	WaitForAllJobsSuccess(t, kc, rmqNamespace, 60, 1)
 	assert.True(t, WaitForScaledJobCount(t, kc, scaledJobName, testNamespace, 8, iterationCount, 1),
 		"job count should be %d after %d iterations", 8, iterationCount)
 
-	RMQPublishMessages(t, rmqNamespace, connectionString, queueName, 4)
+	RMQPublishMessages(t, rmqNamespace, connectionString, queueName, 8)
+	WaitForAllJobsSuccess(t, kc, rmqNamespace, 60, 1)
 	assert.True(t, WaitForScaledJobCount(t, kc, scaledJobName, testNamespace, 10, iterationCount, 1),
 		"job count should be %d after %d iterations", 10, iterationCount)
 }
diff --git a/tests/internals/value_metric_type/value_metric_type_test.go b/tests/internals/value_metric_type/value_metric_type_test.go
@@ -149,8 +149,8 @@ func testScaleByAverageValue(t *testing.T, kc *kubernetes.Clientset, data templa
 
 	// Metric Value = 8, DesiredAverageMetricValue = 2
 	// should scale in to 8/2 = 4 replicas, irrespective of current replicas
-	assert.True(t, WaitForDeploymentReplicaReadyCount(t, kc, deploymentName, testNamespace, 4, 60, 1),
-		"replica count should be 4 after 1 minute")
+	assert.True(t, WaitForDeploymentReplicaReadyCount(t, kc, deploymentName, testNamespace, 4, 60, 3),
+		"replica count should be 4 after 3 minute")
 
 	KubectlDeleteWithTemplate(t, data, "scaledObjectTemplate", scaledObjectTemplate)
 }

diff --git a/tests/run-all.go b/tests/run-all.go
@@ -25,11 +25,11 @@ import (
 )
 
 var (
-	concurrentTests        = 15
+	concurrentTests        = 25
 	regularTestsTimeout    = "20m"
 	regularTestsRetries    = 3
 	sequentialTestsTimeout = "20m"
-	sequentialTestsRetries = 1
+	sequentialTestsRetries = 2
 )
 
 type TestResult struct {

diff --git a/tests/scalers/artemis/artemis_test.go b/tests/scalers/artemis/artemis_test.go
@@ -40,6 +40,7 @@ type templateData struct {
 	SecretName            string
 	ArtemisPasswordBase64 string
 	ArtemisUserBase64     string
+	MessageCount          int
 }
 
 const (
@@ -87,8 +88,8 @@ spec:
     spec:
       containers:
       - name: kedartemis-consumer
-        image: balchu/kedartemis-consumer
-        imagePullPolicy: Always
+        image: ghcr.io/kedacore/tests-artemis
+        args: ["consumer"]
         env:
           - name: ARTEMIS_PASSWORD
             valueFrom:
@@ -100,10 +101,12 @@ spec:
               secretKeyRef:
                 name: {{.SecretName}}
                 key: artemis-username
-          - name: ARTEMIS_HOST
+          - name: ARTEMIS_SERVER_HOST
             value: "artemis-activemq.{{.TestNamespace}}"
-          - name: ARTEMIS_PORT
+          - name: ARTEMIS_SERVER_PORT
             value: "61616"
+          - name: ARTEMIS_MESSAGE_SLEEP_MS
+            value: "70"
 `
 
 	artemisDeploymentTemplate = `apiVersion: apps/v1
@@ -260,7 +263,7 @@ spec:
       managementEndpoint: "artemis-activemq.{{.TestNamespace}}:8161"
       queueName: "test"
       queueLength: "50"
-      activationQueueLength: "1500"
+      activationQueueLength: "5"
       brokerName: "artemis-activemq"
       brokerAddress: "test"
     authenticationRef:
@@ -279,7 +282,8 @@ spec:
     spec:
       containers:
         - name: artemis-producer
-          image: balchu/artemis-producer:0.0.1
+          image: ghcr.io/kedacore/tests-artemis
+          args: ["producer"]
           env:
             - name: ARTEMIS_PASSWORD
               valueFrom:
@@ -295,6 +299,8 @@ spec:
               value: "artemis-activemq.{{.TestNamespace}}"
             - name: ARTEMIS_SERVER_PORT
               value: "61616"
+            - name: ARTEMIS_MESSAGE_COUNT
+              value: "{{.MessageCount}}"
       restartPolicy: Never
   backoffLimit: 4
 `
@@ -321,13 +327,15 @@ func TestArtemisScaler(t *testing.T) {
 
 func testActivation(t *testing.T, kc *kubernetes.Clientset, data templateData) {
 	t.Log("--- testing activation ---")
+	data.MessageCount = 1
 	KubectlReplaceWithTemplate(t, data, "triggerJobTemplate", producerJob)
 
 	AssertReplicaCountNotChangeDuringTimePeriod(t, kc, deploymentName, testNamespace, minReplicaCount, 60)
 }
 
 func testScaleOut(t *testing.T, kc *kubernetes.Clientset, data templateData) {
 	t.Log("--- testing scale out ---")
+	data.MessageCount = 1000
 	KubectlReplaceWithTemplate(t, data, "triggerJobTemplate", producerJob)
 
 	assert.True(t, WaitForDeploymentReplicaReadyCount(t, kc, deploymentName, testNamespace, maxReplicaCount, 60, 3),
@@ -349,6 +357,7 @@ func getTemplateData() (templateData, []Template) {
 			SecretName:            secretName,
 			ArtemisPasswordBase64: base64.StdEncoding.EncodeToString([]byte(artemisPassword)),
 			ArtemisUserBase64:     base64.StdEncoding.EncodeToString([]byte(artemisUser)),
+			MessageCount:          0,
 		}, []Template{
 			{Name: "secretTemplate", Config: secretTemplate},
 			{Name: "triggerAuthenticationTemplate", Config: triggerAuthenticationTemplate},

diff --git a/tests/scalers/azure/azure_event_hub_dapr_wi/azure_event_hub_dapr_wi_test.go b/tests/scalers/azure/azure_event_hub_dapr_wi/azure_event_hub_dapr_wi_test.go
@@ -26,7 +26,7 @@ import (
 var _ = godotenv.Load("../../../.env")
 
 const (
-	testName              = "azure-event-hub-dapr"
+	testName              = "azure-event-hub-dapr-wi"
 	eventhubConsumerGroup = "$Default"
 )
 

diff --git a/tests/scalers/azure/azure_service_bus_queue_regex/azure_service_bus_queue_regex_test.go b/tests/scalers/azure/azure_service_bus_queue_regex/azure_service_bus_queue_regex_test.go
@@ -202,8 +202,8 @@ func testScale(t *testing.T, kc *kubernetes.Clientset, client *azservicebus.Clie
 	// check different aggregation operations
 	data.Operation = "max"
 	KubectlApplyWithTemplate(t, data, "scaledObjectTemplate", scaledObjectTemplate)
-	assert.True(t, WaitForDeploymentReplicaReadyCount(t, kc, deploymentName, testNamespace, 4, 60, 1),
-		"replica count should be 4 after 1 minute")
+	assert.True(t, WaitForDeploymentReplicaReadyCount(t, kc, deploymentName, testNamespace, 4, 60, 3),
+		"replica count should be 4 after 3 minute")
 
 	data.Operation = "avg"
 	KubectlApplyWithTemplate(t, data, "scaledObjectTemplate", scaledObjectTemplate)

diff --git a/tests/scalers/azure/azure_service_bus_topic_regex/azure_service_bus_topic_regex_test.go b/tests/scalers/azure/azure_service_bus_topic_regex/azure_service_bus_topic_regex_test.go
@@ -225,8 +225,8 @@ func testScale(t *testing.T, kc *kubernetes.Clientset, client *azservicebus.Clie
 	// check different aggregation operations
 	data.Operation = "max"
 	KubectlApplyWithTemplate(t, data, "scaledObjectTemplate", scaledObjectTemplate)
-	assert.True(t, WaitForDeploymentReplicaReadyCount(t, kc, deploymentName, testNamespace, 4, 60, 1),
-		"replica count should be 4 after 1 minute")
+	assert.True(t, WaitForDeploymentReplicaReadyCount(t, kc, deploymentName, testNamespace, 4, 60, 3),
+		"replica count should be 4 after 3 minute")
 
 	data.Operation = "avg"
 	KubectlApplyWithTemplate(t, data, "scaledObjectTemplate", scaledObjectTemplate)

diff --git a/tests/scalers/elasticsearch/elasticsearch_test.go b/tests/scalers/elasticsearch/elasticsearch_test.go
@@ -81,7 +81,7 @@ metadata:
   labels:
     app: {{.DeploymentName}}
 spec:
-  replicas: 1
+  replicas: 0
   selector:
     matchLabels:
       app: {{.DeploymentName}}
@@ -397,7 +397,7 @@ func testElasticsearchScaler(t *testing.T, kc *kubernetes.Clientset) {
 	AssertReplicaCountNotChangeDuringTimePeriod(t, kc, deploymentName, testNamespace, minReplicaCount, 60)
 
 	t.Log("--- testing scale out ---")
-	addElements(t, 5)
+	addElements(t, 10)
 
 	assert.True(t, WaitForDeploymentReplicaReadyCount(t, kc, deploymentName, testNamespace, maxReplicaCount, 60, 3),
 		"replica count should be %d after 3 minutes", maxReplicaCount)

diff --git a/tests/scalers/external_scaler_sj/external_scaler_sj_test.go b/tests/scalers/external_scaler_sj/external_scaler_sj_test.go
@@ -9,6 +9,7 @@ import (
 
 	"github.com/joho/godotenv"
 	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
 	"k8s.io/client-go/kubernetes"
 
 	. "github.com/kedacore/keda/v2/tests/helper"
@@ -139,6 +140,9 @@ func TestScaler(t *testing.T) {
 
 	CreateKubernetesResources(t, kc, testNamespace, data, templates)
 
+	require.True(t, WaitForDeploymentReplicaReadyCount(t, kc, scalerName, testNamespace, 1, 60, 1),
+		"replica count should be 1 after 1 minute")
+
 	assert.True(t, WaitForJobCount(t, kc, testNamespace, 0, 60, 1),
 		"job count should be 0 after 1 minute")
 
@@ -184,7 +188,7 @@ func testScaleIn(t *testing.T, kc *kubernetes.Clientset, data templateData) {
 	data.MetricValue = 0
 	KubectlReplaceWithTemplate(t, data, "updateMetricTemplate", updateMetricTemplate)
 
-	assert.True(t, WaitForScaledJobCount(t, kc, scaledJobName, testNamespace, 0, 60, 1),
-		"job count should be 0 after 1 minute")
+	assert.True(t, WaitForScaledJobCount(t, kc, scaledJobName, testNamespace, 0, 120, 1),
+		"job count should be 0 after 2 minute")
 	KubectlDeleteWithTemplate(t, data, "updateMetricTemplate", updateMetricTemplate)
 }
diff --git a/tests/scalers/github_runner/github_runner_test.go b/tests/scalers/github_runner/github_runner_test.go
@@ -313,13 +313,11 @@ func TestScaler(t *testing.T) {
 
 	// test scaling Scaled Job with App
 	KubectlApplyWithTemplate(t, data, "scaledGhaJobTemplate", scaledGhaJobTemplate)
-	// testActivation(t, kc, client)
 	testJobScaleOut(t, kc, client, ghaWorkflowID)
 	testJobScaleIn(t, kc)
 
 	// test scaling Scaled Job
 	KubectlApplyWithTemplate(t, data, "scaledJobTemplate", scaledJobTemplate)
-	// testActivation(t, kc, client)
 	testJobScaleOut(t, kc, client, workflowID)
 	testJobScaleIn(t, kc)