Fix CNS IP demand overcounting by filtering terminal Pods in IPAMv2

cns/ipampool/v2/adapter.go

@@ -31,6 +31,23 @@ func (m *adapter) GetStateSnapshot() cns.IpamPoolMonitorStateSnapshot {
 
 func PodIPDemandListener(ch chan<- int) func([]v1.Pod) {
 	return func(pods []v1.Pod) {
+		// Filter out Pods in terminal phases (Succeeded/Failed) since they no longer
+		// have network sandboxes and don't contribute to IP demand
+		activePods := 0
+		for i := range pods {
+			if pods[i].Status.Phase != v1.PodSucceeded && pods[i].Status.Phase != v1.PodFailed {
+				activePods++
+			}
+		}
+		ch <- activePods
+	}
+}
+
+// PodIPDemandListenerNoFilter counts all pods without client-side filtering.
+// This is used when server-side filtering has already excluded terminal pods.
+func PodIPDemandListenerNoFilter(ch chan<- int) func([]v1.Pod) {
+	return func(pods []v1.Pod) {
+		// All pods are already filtered server-side to exclude terminal phases
 		ch <- len(pods)
 	}
 }

cns/ipampool/v2/adapter_test.go

@@ -0,0 +1,160 @@
+package v2
+
+import (
+	"testing"
+
+	v1 "k8s.io/api/core/v1"
+	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
+)
+
+func TestPodIPDemandListener(t *testing.T) {
+	tests := []struct {
+		name     string
+		pods     []v1.Pod
+		expected int
+	}{
+		{
+			name:     "empty pod list",
+			pods:     []v1.Pod{},
+			expected: 0,
+		},
+		{
+			name: "single running pod",
+			pods: []v1.Pod{
+				{
+					ObjectMeta: metav1.ObjectMeta{Name: "pod1"},
+					Status:     v1.PodStatus{Phase: v1.PodRunning},
+				},
+			},
+			expected: 1,
+		},
+		{
+			name: "multiple running pods",
+			pods: []v1.Pod{
+				{
+					ObjectMeta: metav1.ObjectMeta{Name: "pod1"},
+					Status:     v1.PodStatus{Phase: v1.PodRunning},
+				},
+				{
+					ObjectMeta: metav1.ObjectMeta{Name: "pod2"},
+					Status:     v1.PodStatus{Phase: v1.PodPending},
+				},
+			},
+			expected: 2,
+		},
+		{
+			name: "mix of running and terminal pods - should exclude terminal",
+			pods: []v1.Pod{
+				{
+					ObjectMeta: metav1.ObjectMeta{Name: "pod1"},
+					Status:     v1.PodStatus{Phase: v1.PodRunning},
+				},
+				{
+					ObjectMeta: metav1.ObjectMeta{Name: "pod2"},
+					Status:     v1.PodStatus{Phase: v1.PodSucceeded},
+				},
+				{
+					ObjectMeta: metav1.ObjectMeta{Name: "pod3"},
+					Status:     v1.PodStatus{Phase: v1.PodFailed},
+				},
+				{
+					ObjectMeta: metav1.ObjectMeta{Name: "pod4"},
+					Status:     v1.PodStatus{Phase: v1.PodPending},
+				},
+			},
+			expected: 2, // Only pod1 (Running) and pod4 (Pending) should be counted
+		},
+		{
+			name: "only terminal pods - should count zero",
+			pods: []v1.Pod{
+				{
+					ObjectMeta: metav1.ObjectMeta{Name: "pod1"},
+					Status:     v1.PodStatus{Phase: v1.PodSucceeded},
+				},
+				{
+					ObjectMeta: metav1.ObjectMeta{Name: "pod2"},
+					Status:     v1.PodStatus{Phase: v1.PodFailed},
+				},
+			},
+			expected: 0,
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			ch := make(chan int, 1)
+			listener := PodIPDemandListener(ch)
+
+			listener(tt.pods)
+
+			select {
+			case result := <-ch:
+				if result != tt.expected {
+					t.Errorf("expected %d, got %d", tt.expected, result)
+				}
+			default:
+				t.Error("expected value in channel")
+			}
+		})
+	}
+}
+
+func TestPodIPDemandListenerNoFilter(t *testing.T) {
+	tests := []struct {
+		name     string
+		pods     []v1.Pod
+		expected int
+	}{
+		{
+			name:     "empty pod list",
+			pods:     []v1.Pod{},
+			expected: 0,
+		},
+		{
+			name: "single pod",
+			pods: []v1.Pod{
+				{
+					ObjectMeta: metav1.ObjectMeta{Name: "pod1"},
+					Status:     v1.PodStatus{Phase: v1.PodRunning},
+				},
+			},
+			expected: 1,
+		},
+		{
+			name: "multiple pods - counts all since filtering is done server-side",
+			pods: []v1.Pod{
+				{
+					ObjectMeta: metav1.ObjectMeta{Name: "pod1"},
+					Status:     v1.PodStatus{Phase: v1.PodRunning},
+				},
+				{
+					ObjectMeta: metav1.ObjectMeta{Name: "pod2"},
+					Status:     v1.PodStatus{Phase: v1.PodPending},
+				},
+				{
+					ObjectMeta: metav1.ObjectMeta{Name: "pod3"},
+					Status:     v1.PodStatus{Phase: v1.PodUnknown},
+				},
+			},
+			expected: 3, // All pods counted since server-side filtering already excluded terminal pods
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			ch := make(chan int, 1)
+			listener := PodIPDemandListenerNoFilter(ch)
+
+			listener(tt.pods)
+
+			select {
+			case result := <-ch:
+				if result != tt.expected {
+					t.Errorf("expected %d, got %d", tt.expected, result)
+				}
+			default:
+				t.Error("expected value in channel")
+			}
+		})
+	}
+}

cns/kubecontroller/pod/reconciler.go

@@ -7,6 +7,7 @@ import (
 	"github.com/pkg/errors"
 	"go.uber.org/zap"
 	v1 "k8s.io/api/core/v1"
+	"k8s.io/apimachinery/pkg/fields"
 	ctrl "sigs.k8s.io/controller-runtime"
 	"sigs.k8s.io/controller-runtime/pkg/client"
 	"sigs.k8s.io/controller-runtime/pkg/event"
@@ -50,7 +51,7 @@ type limiter interface {
 	Allow() bool
 }
 
-// NotifierFunc returns a reconcile.Func that lists Pods to get the latest
+// NewNotifierFunc returns a reconcile.Func that lists Pods to get the latest
 // state and notifies listeners of the resulting Pods.
 // listOpts are passed to the client.List call to filter the Pod list.
 // limiter is an optional rate limiter which may be used to limit the
@@ -80,6 +81,54 @@ func (p *watcher) NewNotifierFunc(listOpts *client.ListOptions, limiter limiter,
 	}
 }
 
+// NewPodIPDemandNotifierFunc returns a reconcile.Func optimized for IP demand calculation
+// that uses server-side filtering to exclude terminal Pods (Succeeded/Failed) from the results.
+// This avoids client-side iteration over all pods by using multiple targeted queries.
+func (p *watcher) NewPodIPDemandNotifierFunc(baseListOpts *client.ListOptions, limiter limiter, listeners ...func([]v1.Pod)) reconcile.Func {
+	p.z.Info("adding pod IP demand notifier for listeners", zap.Int("listeners", len(listeners)))
+	return func(ctx context.Context, req ctrl.Request) (ctrl.Result, error) {
+		if !limiter.Allow() {
+			// rate limit exceeded, requeue
+			p.z.Info("rate limit exceeded")
+			return ctrl.Result{Requeue: true}, nil
+		}
+
+		// Create field selectors for active pod phases (exclude Succeeded and Failed)
+		activePhases := []v1.PodPhase{v1.PodRunning, v1.PodPending, v1.PodUnknown}
+		var allActivePods []v1.Pod
+
+		for _, phase := range activePhases {
+			// Clone base list options and add phase filter
+			listOpts := &client.ListOptions{}
+			if baseListOpts.FieldSelector != nil {
+				// Combine existing field selector with phase filter
+				phaseSelector := fields.SelectorFromSet(fields.Set{"status.phase": string(phase)})
+				combinedSelector := fields.AndSelectors(baseListOpts.FieldSelector, phaseSelector)
+				listOpts.FieldSelector = combinedSelector
+			} else {
+				listOpts.FieldSelector = fields.SelectorFromSet(fields.Set{"status.phase": string(phase)})
+			}
+
+			// Copy other options from base
+			listOpts.LabelSelector = baseListOpts.LabelSelector
+			listOpts.Namespace = baseListOpts.Namespace
+			listOpts.Limit = baseListOpts.Limit
+			listOpts.Continue = baseListOpts.Continue
+
+			podList := &v1.PodList{}
+			if err := p.cli.List(ctx, podList, listOpts); err != nil {
+				return ctrl.Result{}, errors.Wrapf(err, "failed to list pods with phase %s", phase)
+			}
+			allActivePods = append(allActivePods, podList.Items...)
+		}
+
+		for _, l := range listeners {
+			l(allActivePods)
+		}
+		return ctrl.Result{}, nil
+	}
+}
+
 var hostNetworkIndexer = client.IndexerFunc(func(o client.Object) []string {
 	pod, ok := o.(*v1.Pod)
 	if !ok {
@@ -88,12 +137,23 @@ var hostNetworkIndexer = client.IndexerFunc(func(o client.Object) []string {
 	return []string{strconv.FormatBool(pod.Spec.HostNetwork)}
 })
 
+var statusPhaseIndexer = client.IndexerFunc(func(o client.Object) []string {
+	pod, ok := o.(*v1.Pod)
+	if !ok {
+		return nil
+	}
+	return []string{string(pod.Status.Phase)}
+})
+
 // SetupWithManager Sets up the reconciler with a new manager, filtering using NodeNetworkConfigFilter on nodeName.
 func (p *watcher) SetupWithManager(ctx context.Context, mgr ctrl.Manager) error {
 	p.cli = mgr.GetClient()
 	if err := mgr.GetFieldIndexer().IndexField(ctx, &v1.Pod{}, "spec.hostNetwork", hostNetworkIndexer); err != nil {
 		return errors.Wrap(err, "failed to set up hostNetwork indexer")
 	}
+	if err := mgr.GetFieldIndexer().IndexField(ctx, &v1.Pod{}, "status.phase", statusPhaseIndexer); err != nil {
+		return errors.Wrap(err, "failed to set up status.phase indexer")
+	}
 	if err := ctrl.NewControllerManagedBy(mgr).
 		For(&v1.Pod{}).
 		WithEventFilter(predicate.Funcs{ // we only want create/delete events

cns/service/main.go

@@ -1576,10 +1576,16 @@ func InitializeCRDState(ctx context.Context, httpRestService cns.HTTPService, cn
 	if cnsconfig.WatchPods {
 		pw := podctrl.New(z)
 		if cnsconfig.EnableIPAMv2 {
-			hostNetworkListOpt := &client.ListOptions{FieldSelector: fields.SelectorFromSet(fields.Set{"spec.hostNetwork": "false"})} // filter only podsubnet pods
+			// Create field selector to filter only podsubnet pods
+			fieldSet := fields.Set{
+				"spec.hostNetwork": "false",
+			}
+			hostNetworkSelector := fields.SelectorFromSet(fieldSet)
+			baseListOpts := &client.ListOptions{FieldSelector: hostNetworkSelector}
 			// don't relist pods more than every 500ms
 			limit := rate.NewLimiter(rate.Every(500*time.Millisecond), 1) //nolint:gomnd // clearly 500ms
-			pw.With(pw.NewNotifierFunc(hostNetworkListOpt, limit, ipampoolv2.PodIPDemandListener(ipDemandCh)))
+			// Use specialized notifier that filters terminal pods server-side
+			pw.With(pw.NewPodIPDemandNotifierFunc(baseListOpts, limit, ipampoolv2.PodIPDemandListenerNoFilter(ipDemandCh)))
 		}
 		if err := pw.SetupWithManager(ctx, manager); err != nil {
 			return errors.Wrapf(err, "failed to setup pod watcher with manager")