fix(e2e): wait for leader election & measure timing for better monitoring

TestClusterExtensionAfterOLMUpgrade was failing due to increased leader
election timeouts, causing reconciliation checks to run before leadership
was acquired.

This fix ensures the test explicitly waits for leader election logs
(`"successfully acquired lease"`) before verifying reconciliation.

Additionally, the test now measures and logs the leader election duration
to help monitor election timing.

Author: camilamacedo86

catalogd/cmd/catalogd/main.go

@@ -235,9 +235,9 @@ func main() {
 		LeaderElectionID:       "catalogd-operator-lock",
 		// Recommended Leader Election values
 		// https://github.com/openshift/enhancements/blob/61581dcd985130357d6e4b0e72b87ee35394bf6e/CONVENTIONS.md#handling-kube-apiserver-disruption
-		LeaseDuration: ptr.To(137 * time.Second),
-		RenewDeadline: ptr.To(107 * time.Second),
-		RetryPeriod:   ptr.To(26 * time.Second),
+		LeaseDuration: ptr.To(137 * time.Second), // Default: 15s
+		RenewDeadline: ptr.To(107 * time.Second), // Default: 10s
+		RetryPeriod:   ptr.To(26 * time.Second),  // Default: 2s
 
 		WebhookServer: webhookServer,
 		Cache:         cacheOptions,

cmd/operator-controller/main.go

@@ -235,9 +235,9 @@ func main() {
 		LeaderElectionID:       "9c4404e7.operatorframework.io",
 		// Recommended Leader Election values
 		// https://github.com/openshift/enhancements/blob/61581dcd985130357d6e4b0e72b87ee35394bf6e/CONVENTIONS.md#handling-kube-apiserver-disruption
-		LeaseDuration: ptr.To(137 * time.Second),
-		RenewDeadline: ptr.To(107 * time.Second),
-		RetryPeriod:   ptr.To(26 * time.Second),
+		LeaseDuration: ptr.To(137 * time.Second), // Default: 15s
+		RenewDeadline: ptr.To(107 * time.Second), // Default: 10s
+		RetryPeriod:   ptr.To(26 * time.Second),  // Default: 2s
 
 		Cache: cacheOptions,
 		// LeaderElectionReleaseOnCancel defines if the leader should step down voluntarily

test/upgrade-e2e/post_upgrade_test.go

@@ -40,6 +40,20 @@ func TestClusterExtensionAfterOLMUpgrade(t *testing.T) {
 	t.Log("Wait for operator-controller deployment to be ready")
 	managerPod := waitForDeployment(t, ctx, "operator-controller-controller-manager")
 
+	t.Log("Start measuring leader election time")
+	// Average case is under 1 minute but in the worst case: (previous leader crashed)
+	// we could have LeaseDuration (137s) + RetryPeriod (26s) +/- 163s
+	leaderStartTime := time.Now()
+	leaderElectionCtx, leaderCancel := context.WithTimeout(ctx, 3*time.Minute)
+	defer leaderCancel()
+
+	leaderSubstrings := []string{"successfully acquired lease"}
+	leaderElected, err := watchPodLogsForSubstring(leaderElectionCtx, managerPod, "manager", leaderSubstrings...)
+	require.NoError(t, err)
+	require.True(t, leaderElected)
+	leaderElectionDuration := time.Since(leaderStartTime)
+	t.Logf("Leader election took %v seconds", leaderElectionDuration.Seconds())
+
 	t.Log("Reading logs to make sure that ClusterExtension was reconciled by operator-controller before we update it")
 	// Make sure that after we upgrade OLM itself we can still reconcile old objects without any changes
 	logCtx, cancel := context.WithTimeout(ctx, time.Minute)