koord-manager: ClusterColocationProfile support grayscale

apis/config/v1alpha1/cluster_colocation_profile_types.go

@@ -19,6 +19,7 @@ package v1alpha1
 import (
 	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
 	"k8s.io/apimachinery/pkg/runtime"
+	"k8s.io/apimachinery/pkg/util/intstr"
 )
 
 // EDIT THIS FILE!  THIS IS SCAFFOLDING FOR YOU TO OWN!
@@ -38,6 +39,10 @@ type ClusterColocationProfileSpec struct {
 	// +optional
 	Selector *metav1.LabelSelector `json:"selector,omitempty"`
 
+	// Probability indicates profile will make effect with a probability.
+	// +optional
+	Probability *intstr.IntOrString `json:"probability,omitempty"`
+
 	// QoSClass describes the type of Koordinator QoS that the Pod is running.
 	// The value will be injected into Pod as label koordinator.sh/qosClass.
 	// Options are LSE/LSR/LS/BE/SYSTEM.

config/crd/bases/config.koordinator.sh_clustercolocationprofiles.yaml

@@ -119,6 +119,13 @@ spec:
                 - koord-batch
                 - koord-free
                 type: string
+              probability:
+                anyOf:
+                - type: integer
+                - type: string
+                description: Probability indicates profile will make effect with a
+                  probability.
+                x-kubernetes-int-or-string: true
               qosClass:
                 description: QoSClass describes the type of Koordinator QoS that the
                   Pod is running. The value will be injected into Pod as label koordinator.sh/qosClass.

pkg/webhook/pod/mutating/cluster_colocation_profile.go

@@ -20,6 +20,7 @@ import (
 	"context"
 	"encoding/json"
 	"fmt"
+	"math/rand"
 
 	admissionv1 "k8s.io/api/admission/v1"
 	corev1 "k8s.io/api/core/v1"
@@ -28,6 +29,7 @@ import (
 	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
 	"k8s.io/apimachinery/pkg/labels"
 	"k8s.io/apimachinery/pkg/types"
+	"k8s.io/apimachinery/pkg/util/intstr"
 	"k8s.io/apimachinery/pkg/util/strategicpatch"
 	"k8s.io/klog/v2"
 	"k8s.io/utils/pointer"
@@ -39,6 +41,10 @@ import (
 	utilclient "github.com/koordinator-sh/koordinator/pkg/util/client"
 )
 
+var (
+	randIntnFn = rand.Intn
+)
+
 // +kubebuilder:rbac:groups=core,resources=namespaces,verbs=get;list;watch
 // +kubebuilder:rbac:groups=config.koordinator.sh,resources=clustercolocationprofiles,verbs=get;list;watch
 
@@ -79,7 +85,15 @@ func (h *PodMutatingHandler) clusterColocationProfileMutatingPod(ctx context.Con
 	}
 
 	for _, profile := range matchedProfiles {
-		err := h.doMutateByColocationProfile(ctx, pod, profile)
+		skip, err := shouldSkipProfile(profile)
+		if err != nil {
+			return err
+		}
+		if skip {
+			klog.V(4).Infof("skip mutate Pod %s/%s by clusterColocationProfile %s", pod.Namespace, pod.Name, profile.Name)
+			continue
+		}
+		err = h.doMutateByColocationProfile(ctx, pod, profile)
 		if err != nil {
 			return err
 		}
@@ -125,6 +139,18 @@ func (h *PodMutatingHandler) matchObjectSelector(pod, oldPod *corev1.Pod, object
 	return matched, nil
 }
 
+func shouldSkipProfile(profile *configv1alpha1.ClusterColocationProfile) (bool, error) {
+	percent := 100
+	if profile.Spec.Probability != nil {
+		var err error
+		percent, err = intstr.GetScaledValueFromIntOrPercent(profile.Spec.Probability, 100, false)
+		if err != nil {
+			return false, err
+		}
+	}
+	return percent == 0 || (percent != 100 && randIntnFn(100) > percent), nil
+}
+
 func (h *PodMutatingHandler) doMutateByColocationProfile(ctx context.Context, pod *corev1.Pod, profile *configv1alpha1.ClusterColocationProfile) error {
 	if len(profile.Spec.Labels) > 0 {
 		if pod.Labels == nil {