Update restart events to put resource in regarding (strimzi#11352)

katheris · web-flow · commit 05dfb2e56545 · 2025-04-22T13:10:06.000+02:00
Signed-off-by: Katherine Stanley &lt;11195226+katheris@users.noreply.github.com&gt;
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -23,6 +23,7 @@
 * The `ContinueReconciliationOnManualRollingUpdateFailure` feature gate moves to GA stage and is permanently enabled without the possibility to disable it.
 * Update OAuth library to 0.16.2.
 * Update HTTP bridge to 0.32.0.
+* Kubernetes events emitted during a Pod restart updated to have the Kafka resource as the `regardingObject` and the Pod in the `related` field.
 
 ### Major changes, deprecations and removals
 
@@ -45,6 +46,8 @@
   Please use the template section to configure additional volumes instead.
 * Kafka 4.0 and newer is using Log4j2 for logging instead of Reload4j/Log4j1.
   If you have any custom logging configuration, you might need to update it during the upgrade to Kafka 4.0.
+* Kubernetes events for Pod restarts no longer have the Pod as the `regardingObject`.
+  If you are using `regardingObject` as a `field-selector` for listing events you must update the selector to specify the Kafka resource instead.
 
 ## 0.45.0
 
diff --git a/cluster-operator/src/main/java/io/strimzi/operator/cluster/operator/resource/KafkaRoller.java b/cluster-operator/src/main/java/io/strimzi/operator/cluster/operator/resource/KafkaRoller.java
@@ -903,7 +903,7 @@ private static <T, E extends Exception> T await(Future<T> future, long timeout,
     protected Future<Void> restart(Pod pod, RestartContext restartContext) {
         return podOperations.restart(reconciliation, pod, operationTimeoutMs)
                              .onComplete(i -> vertx.executeBlocking(() -> {
-                                 eventsPublisher.publishRestartEvents(pod, restartContext.restartReasons);
+                                 eventsPublisher.publishRestartEvents(reconciliation, pod, restartContext.restartReasons);
                                  return null;
                              }));
     }
diff --git a/cluster-operator/src/main/java/io/strimzi/operator/cluster/operator/resource/events/KubernetesRestartEventPublisher.java b/cluster-operator/src/main/java/io/strimzi/operator/cluster/operator/resource/events/KubernetesRestartEventPublisher.java
@@ -12,6 +12,7 @@
 import io.fabric8.kubernetes.client.KubernetesClient;
 import io.strimzi.operator.cluster.model.RestartReason;
 import io.strimzi.operator.cluster.model.RestartReasons;
+import io.strimzi.operator.common.Reconciliation;
 import org.apache.logging.log4j.LogManager;
 import org.apache.logging.log4j.Logger;
 
@@ -57,23 +58,25 @@ protected KubernetesRestartEventPublisher(KubernetesClient client, String operat
     }
 
     /**
-     * Publishes an Kubernetes Event about Pod restart
+     * Publishes a Kubernetes Event about Pod restart
      *
-     * @param pod       Pod which is restarted
-     * @param reasons   Reasons for the restart
+     * @param reconciliation    Reconciliation marker
+     * @param pod               Pod which is restarted
+     * @param reasons           Reasons for the restart
      */
-    public void publishRestartEvents(Pod pod, RestartReasons reasons) {
+    public void publishRestartEvents(Reconciliation reconciliation, Pod pod, RestartReasons reasons) {
         MicroTime k8sEventTime = new MicroTime(K8S_MICROTIME.format(ZonedDateTime.now(clock)));
+        ObjectReference resourceReference = createResourceReference(reconciliation);
         ObjectReference podReference = createPodReference(pod);
 
         try {
             for (RestartReason reason : reasons) {
-                String note = maybeTruncated(reasons.getNoteFor(reason));
+                String note = maybeTruncated("Rolling Pod " + pod.getMetadata().getName() + " due to " + reasons.getNoteFor(reason));
                 String type = "Normal";
                 String k8sFormattedReason = reason.pascalCased();
-                LOG.debug("Publishing K8s event, time {}, type, {}, reason, {}, note, {}, pod, {}",
-                        k8sEventTime, type, k8sFormattedReason, note, podReference);
-                publishEvent(k8sEventTime, podReference, k8sFormattedReason, type, note);
+                LOG.debug("Publishing K8s event, time={}, type={}, reason={}, note={}, resource={}, pod={}",
+                        k8sEventTime, type, k8sFormattedReason, note, resourceReference, podReference);
+                publishEvent(k8sEventTime, resourceReference, podReference, k8sFormattedReason, type, note);
             }
         } catch (Exception e) {
             LOG.error("Exception on K8s event publication", e);
@@ -83,13 +86,14 @@ public void publishRestartEvents(Pod pod, RestartReasons reasons) {
     /**
      * Publish a Kubernetes Event referring to certain KafkaRoller pod action
      *
-     * @param eventTime    - Microtime to use for event
-     * @param podReference - ObjectReference pointing to rolled pod
-     * @param reason       - reason the pod is being rolled
-     * @param type         - the type of K8s event "Normal", or "Warning"
-     * @param note         - the note to attach to the event
+     * @param eventTime         - Microtime to use for event
+     * @param resourceReference - ObjectReference pointing to the owning resource
+     * @param podReference      - ObjectReference pointing to rolled pod
+     * @param reason            - reason the pod is being rolled
+     * @param type              - the type of K8s event "Normal", or "Warning"
+     * @param note              - the note to attach to the event
      */
-    protected void publishEvent(MicroTime eventTime, ObjectReference podReference, String reason, String type, String note) {
+    protected void publishEvent(MicroTime eventTime, ObjectReference resourceReference, ObjectReference podReference, String reason, String type, String note) {
         EventBuilder builder = new EventBuilder();
 
         builder.withNewMetadata()
@@ -98,7 +102,8 @@ protected void publishEvent(MicroTime eventTime, ObjectReference podReference, S
                 .withAction(ACTION)
                 .withReportingController(CONTROLLER)
                 .withReportingInstance(operatorName)
-                .withRegarding(podReference)
+                .withRegarding(resourceReference)
+                .withRelated(podReference)
                 .withReason(reason)
                 .withType(type)
                 .withEventTime(eventTime)
@@ -114,6 +119,14 @@ ObjectReference createPodReference(Pod pod) {
                                            .build();
     }
 
+    ObjectReference createResourceReference(Reconciliation reconciliation) {
+        return new ObjectReferenceBuilder()
+                .withKind(reconciliation.kind())
+                .withNamespace(reconciliation.namespace())
+                .withName(reconciliation.name())
+                .build();
+    }
+
 
     /**
      * While the core event API doesn't set a limit on note sizes, events.k8s.io/v1beta1 and v1 do, which is 1kB.
diff --git a/cluster-operator/src/test/java/io/strimzi/operator/cluster/operator/resource/events/KubernetesRestartEventPublisherIT.java b/cluster-operator/src/test/java/io/strimzi/operator/cluster/operator/resource/events/KubernetesRestartEventPublisherIT.java
@@ -14,8 +14,10 @@
 import io.fabric8.kubernetes.api.model.events.v1.Event;
 import io.fabric8.kubernetes.client.KubernetesClient;
 import io.fabric8.kubernetes.client.KubernetesClientBuilder;
+import io.strimzi.api.kafka.model.kafka.Kafka;
 import io.strimzi.operator.cluster.model.RestartReason;
 import io.strimzi.operator.cluster.model.RestartReasons;
+import io.strimzi.operator.common.Reconciliation;
 import io.strimzi.test.TestUtils;
 import org.junit.jupiter.api.AfterAll;
 import org.junit.jupiter.api.AfterEach;
@@ -68,7 +70,8 @@ void teardown() {
     @Test
     void eventPublicationSucceeds() {
         KubernetesRestartEventPublisher publisher = new KubernetesRestartEventPublisher(client, "op") { };
-        publisher.publishRestartEvents(pod, RestartReasons.of(RestartReason.CLUSTER_CA_CERT_KEY_REPLACED).add(RestartReason.FILE_SYSTEM_RESIZE_NEEDED));
+        Reconciliation reconciliation = new Reconciliation("test", Kafka.RESOURCE_KIND, TEST_NAMESPACE, "test");
+        publisher.publishRestartEvents(reconciliation, pod, RestartReasons.of(RestartReason.CLUSTER_CA_CERT_KEY_REPLACED).add(RestartReason.FILE_SYSTEM_RESIZE_NEEDED));
 
         ListOptions strimziEventsOnly = new ListOptionsBuilder()
                 .withFieldSelector("reportingController=" + KubernetesRestartEventPublisher.CONTROLLER)
@@ -79,7 +82,16 @@ void eventPublicationSucceeds() {
 
         Event exemplar = items.get(0);
         assertThat(exemplar.getAction(), is(KubernetesRestartEventPublisher.ACTION));
-        assertThat(exemplar.getRegarding(), is(referenceFromPod(pod)));
+        assertThat(exemplar.getRegarding(), is(referenceFromReconciliation(reconciliation)));
+        assertThat(exemplar.getRelated(), is(referenceFromPod(pod)));
+    }
+
+    static ObjectReference referenceFromReconciliation(Reconciliation reconciliation) {
+        return new ObjectReferenceBuilder()
+                .withKind(reconciliation.kind())
+                .withNamespace(reconciliation.namespace())
+                .withName(reconciliation.name())
+                .build();
     }
 
     static ObjectReference referenceFromPod(Pod pod) {
diff --git a/cluster-operator/src/test/java/io/strimzi/operator/cluster/operator/resource/events/KubernetesRestartEventPublisherTest.java b/cluster-operator/src/test/java/io/strimzi/operator/cluster/operator/resource/events/KubernetesRestartEventPublisherTest.java
@@ -18,8 +18,10 @@
 import io.fabric8.kubernetes.client.dsl.NonNamespaceOperation;
 import io.fabric8.kubernetes.client.dsl.Resource;
 import io.fabric8.kubernetes.client.dsl.V1EventingAPIGroupDSL;
+import io.strimzi.api.kafka.model.kafka.Kafka;
 import io.strimzi.operator.cluster.model.RestartReason;
 import io.strimzi.operator.cluster.model.RestartReasons;
+import io.strimzi.operator.common.Reconciliation;
 import org.junit.jupiter.api.Assertions;
 import org.junit.jupiter.api.BeforeEach;
 import org.junit.jupiter.api.Test;
@@ -42,6 +44,7 @@
 
 class KubernetesRestartEventPublisherTest {
     private final static String NAMESPACE = "test-ns";
+    private final static String CLUSTER_NAME = "example-cluster";
     private final static String POD_NAME = "example-pod";
 
     private KubernetesRestartEventPublisher publisher;
@@ -51,7 +54,7 @@ void setup() {
         KubernetesClient client = mock(KubernetesClient.class);
         publisher = new KubernetesRestartEventPublisher(client, "op") {
             @Override
-            protected void publishEvent(MicroTime eventTime, ObjectReference podReference, String reason, String type, String note) {
+            protected void publishEvent(MicroTime eventTime, ObjectReference resourceReference, ObjectReference podReference, String reason, String type, String note) {
             }
         };
     }
@@ -72,6 +75,22 @@ void testObjectReferenceFromPod() {
         assertThat(podRef.getKind(), is("Pod"));
     }
 
+    @Test
+    void testObjectReferenceFromReconciliation() {
+        Reconciliation reconciliation = new Reconciliation(
+                "test-trigger",
+                Kafka.RESOURCE_KIND,
+                NAMESPACE,
+                CLUSTER_NAME
+        );
+
+        ObjectReference clusterRef = publisher.createResourceReference(reconciliation);
+
+        assertThat(clusterRef.getName(), is(CLUSTER_NAME));
+        assertThat(clusterRef.getNamespace(), is(NAMESPACE));
+        assertThat(clusterRef.getKind(), is(Kafka.RESOURCE_KIND));
+    }
+
     @Test
     void testTruncation() {
         String underOneThousandBytes = "1".repeat(999);
@@ -107,7 +126,7 @@ void testOneEventPublishedPerReason() {
         Set<String> capturedReasons = new HashSet<>();
         KubernetesRestartEventPublisher capturingPublisher = new KubernetesRestartEventPublisher(client, "op") {
             @Override
-            protected void publishEvent(MicroTime eventTime, ObjectReference podReference, String reason, String type, String note) {
+            protected void publishEvent(MicroTime eventTime, ObjectReference resourceReference, ObjectReference podReference, String reason, String type, String note) {
                 capturedReasons.add(reason);
             }
         };
@@ -117,7 +136,7 @@ protected void publishEvent(MicroTime eventTime, ObjectReference podReference, S
         RestartReasons reasons = new RestartReasons().add(RestartReason.FILE_SYSTEM_RESIZE_NEEDED)
                                                      .add(RestartReason.CLUSTER_CA_CERT_KEY_REPLACED);
 
-        capturingPublisher.publishRestartEvents(mockPod, reasons);
+        capturingPublisher.publishRestartEvents(Reconciliation.DUMMY_RECONCILIATION, mockPod, reasons);
 
         assertThat(capturedReasons, is(expectedReasons));
 
@@ -159,22 +178,26 @@ void testPopulatesExpectedFields() {
         KubernetesRestartEventPublisher eventPublisher = new KubernetesRestartEventPublisher(client, "cluster-operator-id", clock);
 
         RestartReasons reasons = new RestartReasons().add(RestartReason.FILE_SYSTEM_RESIZE_NEEDED);
-        eventPublisher.publishRestartEvents(pod, reasons);
+        eventPublisher.publishRestartEvents(new Reconciliation("test", Kafka.RESOURCE_KIND, NAMESPACE, CLUSTER_NAME), pod, reasons);
 
         verify(mockEventResource, times(1)).create();
 
         Event publishedEvent = eventCaptor.getValue();
-        assertThat(publishedEvent.getRegarding().getKind(), is("Pod"));
-        assertThat(publishedEvent.getRegarding().getName(), is(POD_NAME));
+        assertThat(publishedEvent.getRegarding().getKind(), is(Kafka.RESOURCE_KIND));
+        assertThat(publishedEvent.getRegarding().getName(), is(CLUSTER_NAME));
         assertThat(publishedEvent.getRegarding().getNamespace(), is(NAMESPACE));
 
+        assertThat(publishedEvent.getRelated().getKind(), is("Pod"));
+        assertThat(publishedEvent.getRelated().getName(), is(POD_NAME));
+        assertThat(publishedEvent.getRelated().getNamespace(), is(NAMESPACE));
+
         assertThat(publishedEvent.getReportingController(), is("strimzi.io/cluster-operator"));
         assertThat(publishedEvent.getReportingInstance(), is("cluster-operator-id"));
 
         assertThat(publishedEvent.getReason(), is("FileSystemResizeNeeded"));
         assertThat(publishedEvent.getAction(), is("StrimziInitiatedPodRestart"));
         assertThat(publishedEvent.getType(), is("Normal"));
-        assertThat(publishedEvent.getNote(), is(RestartReason.FILE_SYSTEM_RESIZE_NEEDED.getDefaultNote()));
+        assertThat(publishedEvent.getNote(), is("Rolling Pod " + POD_NAME + " due to " + RestartReason.FILE_SYSTEM_RESIZE_NEEDED.getDefaultNote()));
         assertThat(publishedEvent.getEventTime().getTime(), is("2020-10-11T00:00:00.000000Z"));
 
     }
diff --git a/cluster-operator/src/test/java/io/strimzi/operator/cluster/operator/resource/events/KubernetesRestartEventsMockTest.java b/cluster-operator/src/test/java/io/strimzi/operator/cluster/operator/resource/events/KubernetesRestartEventsMockTest.java
@@ -555,7 +555,9 @@ private <T> Handler<AsyncResult<T>> verifyEventPublished(RestartReason expectedR
             }
 
             Event restartEvent = maybeEvent.get();
-            assertThat(restartEvent.getRegarding().getName(), is(kafkaPod().getMetadata().getName()));
+            assertThat(restartEvent.getRelated().getName(), is(kafkaPod().getMetadata().getName()));
+            assertThat(restartEvent.getRegarding().getName(), is(CLUSTER_NAME));
+            assertThat(restartEvent.getRegarding().getKind(), is(Kafka.RESOURCE_KIND));
             context.completeNow();
         }));
     }
diff --git a/documentation/modules/deploying/proc-operator-restart-events.adoc b/documentation/modules/deploying/proc-operator-restart-events.adoc
@@ -28,10 +28,10 @@ kubectl -n kafka get events --field-selector reportingController=strimzi.io/clus
 .Example showing events returned
 [source,shell]
 ----
-LAST SEEN   TYPE     REASON                   OBJECT                        MESSAGE
-2m          Normal   CaCertRenewed            pod/strimzi-cluster-kafka-0   CA certificate renewed
-58m         Normal   PodForceRestartOnError   pod/strimzi-cluster-kafka-1   Pod needs to be forcibly restarted due to an error
-5m47s       Normal   ManualRollingUpdate      pod/strimzi-cluster-kafka-2   Pod was manually annotated to be rolled
+LAST SEEN   TYPE     REASON                   OBJECT                  MESSAGE
+2m          Normal   CaCertRenewed            kafka/strimzi-cluster   Rolling Pod strimzi-cluster-kafka-0 due to CA certificate renewed
+58m         Normal   PodForceRestartOnError   kafka/strimzi-cluster   Rolling Pod strimzi-cluster-kafka-1 due to Pod needs to be forcibly restarted due to an error
+5m47s       Normal   ManualRollingUpdate      kafka/strimzi-cluster   Rolling Pod strimzi-cluster-kafka-2 due to Pod was manually annotated to be rolled
 ----
 +
 You can also specify a `reason` or other `field-selector` options to constrain the events returned.
@@ -60,12 +60,12 @@ items:
   eventTime: "2022-05-13T00:22:34.168086Z"
   firstTimestamp: null
   involvedObject:
-      kind: Pod
-      name: strimzi-cluster-kafka-1
+      kind: Kafka
+      name: strimzi-cluster
       namespace: kafka
   kind: Event
   lastTimestamp: null
-  message: Pod needs to be forcibly restarted due to an error
+  message: Rolling Pod strimzi-cluster-kafka-1 due to Pod needs to be forcibly restarted due to an error
   metadata:
       creationTimestamp: "2022-05-13T00:22:34Z"
       generateName: strimzi-event
@@ -74,6 +74,10 @@ items:
       resourceVersion: "432961"
       uid: 29fcdb9e-f2cf-4c95-a165-a5efcd48edfc
   reason: PodForceRestartOnError
+  related:
+      kind: Pod
+      name: strimzi-cluster-kafka-1
+      namespace: kafka
   reportingController: strimzi.io/cluster-operator
   reportingInstance: strimzi-cluster-operator-6458cfb4c6-6bpdp
   source: {}
diff --git a/documentation/modules/deploying/ref-operator-restart-events-fields.adoc b/documentation/modules/deploying/ref-operator-restart-events-fields.adoc
@@ -10,11 +10,11 @@ When checking restart events from the command line, you can specify a `field-sel
 
 The following fields are available when filtering events with `field-selector`.
 
-`regardingObject.kind`:: The object that was restarted, and for restart events, the kind is always `Pod`.
-`regarding.namespace`:: The namespace that the pod belongs to.
-`regardingObject.name`:: The pod's name, for example, `strimzi-cluster-kafka-0`.
-`regardingObject.uid`:: The unique ID of the pod.
-`reason`:: The reason the pod was restarted, for example, `JbodVolumesChanged`.
+`regardingObject.kind`:: The resource that owns the Pod being restarted, and for restart events, the kind is always `Kafka`.
+`regarding.namespace`:: The namespace that the resource belongs to.
+`regardingObject.name`:: The resource's name, for example, `strimzi-cluster`.
+`regardingObject.uid`:: The unique ID of the resource.
+`reason`:: The reason the Pod was restarted, for example, `JbodVolumesChanged`.
 `reportingController`:: The reporting component is always `strimzi.io/cluster-operator` for Strimzi restart events.
 `source`:: `source` is an older version of `reportingController`. The reporting component is always `strimzi.io/cluster-operator` for Strimzi restart events.
 `type`:: The event type, which is either `Warning` or `Normal`. For Strimzi restart events, the type is `Normal`.

Original file line number	Diff line number	Diff line change
`@@ -903,7 +903,7 @@ private static <T, E extends Exception> T await(Future<T> future, long timeout,`
`903`	`903`	`protected Future<Void> restart(Pod pod, RestartContext restartContext) {`
`904`	`904`	`return podOperations.restart(reconciliation, pod, operationTimeoutMs)`
`905`	`905`	`.onComplete(i -> vertx.executeBlocking(() -> {`
`906`		`- eventsPublisher.publishRestartEvents(pod, restartContext.restartReasons);`
	`906`	`+ eventsPublisher.publishRestartEvents(reconciliation, pod, restartContext.restartReasons);`
`907`	`907`	`return null;`
`908`	`908`	`}));`
`909`	`909`	`}`
Original file line number	Diff line number	Diff line change
`@@ -555,7 +555,9 @@ private <T> Handler<AsyncResult<T>> verifyEventPublished(RestartReason expectedR`
`555`	`555`	`}`
`556`	`556`
`557`	`557`	`Event restartEvent = maybeEvent.get();`
`558`		`- assertThat(restartEvent.getRegarding().getName(), is(kafkaPod().getMetadata().getName()));`
	`558`	`+ assertThat(restartEvent.getRelated().getName(), is(kafkaPod().getMetadata().getName()));`
	`559`	`+ assertThat(restartEvent.getRegarding().getName(), is(CLUSTER_NAME));`
	`560`	`+ assertThat(restartEvent.getRegarding().getKind(), is(Kafka.RESOURCE_KIND));`
`559`	`561`	`context.completeNow();`
`560`	`562`	`}));`
`561`	`563`	`}`