fix(integration): Use kubectl wait for etcd readiness in OLM test

The 'validateOlmAddon' integration test was using a custom retry loop with 'retry.Expo' to check if the etcd operator's CSV had been successfully installed. This could lead to intermittent failures (flakes) in CI.

This change replaces the custom Go retry loop with a single, more idiomatic 'kubectl wait' command. The command waits for the CSV's status phase to become 'Succeeded', using 'kubectl''s built-in polling mechanism.

This makes the test more robust and less prone to flakiness.

Author: Gemini CLI User

test/integration/addons_test.go

@@ -513,23 +513,9 @@ func validateOlmAddon(ctx context.Context, t *testing.T, profile string) {
 		t.Logf("etcd operator installation with %s failed: %v", rr.Command(), err)
 	}
 
-	want := "Succeeded"
-	checkOperatorInstalled := func() error {
-		rr, err := Run(t, exec.CommandContext(ctx, "kubectl", "--context", profile, "get", "csv", "-n", "my-etcd"))
-		if err != nil {
-			return err
-		}
-		if rr.Stderr.String() != "" {
-			t.Logf("%v: unexpected stderr: %s", rr.Command(), rr.Stderr)
-		}
-		if !strings.Contains(rr.Stdout.String(), want) {
-			return fmt.Errorf("%v stdout = %q, want %q", rr.Command(), rr.Stdout, want)
-		}
-		return nil
-	}
 	// Operator installation takes a while
-	if err := retry.Expo(checkOperatorInstalled, time.Second*3, Minutes(10)); err != nil {
-		t.Errorf("failed checking operator installed: %v", err.Error())
+	if _, err := Run(t, exec.CommandContext(ctx, "kubectl", "--context", profile, "wait", "csv", "--all", "-n", "my-etcd", "--for=jsonpath={.status.phase}=Succeeded", "--timeout=10m")); err != nil {
+		t.Errorf("failed waiting for etcd operator to be installed: %v", err)
 	}
 }