feat(watchdog): 主机侧巡检定时器按异常邮件通知平台所有者，operator 增加 phase 与 build_info 指标、卡死存活探针与告警规则

internal/operator/gauge_test.go

@@ -48,3 +48,30 @@ func TestGaugeSyncerPublishesFleetStateFromCluster(t *testing.T) {
 		t.Errorf("servers_total{state=Stopped} = %v, want 2", got)
 	}
 }
+
+// TestGaugeSyncerPublishesServerPhases: every server gets one felis_server_phase
+// series carrying its system role, observed phase (Unknown before the first
+// report) and desired state.
+func TestGaugeSyncerPublishesServerPhases(t *testing.T) {
+	login := gaugeServer("login", v1alpha1.DesiredRunning)
+	login.Labels = map[string]string{v1alpha1.LabelSystemRole: "login"}
+	login.Status.Phase = v1alpha1.PhaseFailed
+	fresh := gaugeServer("fresh", "")
+	c := fake.NewClientBuilder().WithScheme(newScheme(t)).WithObjects(login, fresh).Build()
+
+	if err := (&operator.GaugeSyncer{Client: c}).SyncOnce(context.Background()); err != nil {
+		t.Fatalf("SyncOnce: %v", err)
+	}
+	defer metrics.ServerPhase.Reset()
+	if n := testutil.CollectAndCount(metrics.ServerPhase); n != 2 {
+		t.Fatalf("felis_server_phase series = %d, want 2", n)
+	}
+	for _, labels := range [][]string{
+		{"login", "login", "Failed", "Running"},
+		{"fresh", "", "Unknown", "Stopped"},
+	} {
+		if got := testutil.ToFloat64(metrics.ServerPhase.WithLabelValues(labels...)); got != 1 {
+			t.Errorf("felis_server_phase%v = %v, want 1", labels, got)
+		}
+	}
+}