feat: add sandbox_creation_latency metric

Adds a new histogram metric to track the time it takes for a sandbox
to become ready after it has been created.

Fixes #123

Author: barney-s

controllers/sandbox_controller.go

@@ -22,6 +22,7 @@ import (
 	"reflect"
 	"time"
 
+	"github.com/prometheus/client_golang/prometheus"
 	corev1 "k8s.io/api/core/v1"
 	k8serrors "k8s.io/apimachinery/pkg/api/errors"
 	"k8s.io/apimachinery/pkg/api/meta"
@@ -35,6 +36,7 @@ import (
 	"sigs.k8s.io/controller-runtime/pkg/client"
 	"sigs.k8s.io/controller-runtime/pkg/handler"
 	"sigs.k8s.io/controller-runtime/pkg/log"
+	"sigs.k8s.io/controller-runtime/pkg/metrics"
 	"sigs.k8s.io/controller-runtime/pkg/predicate"
 
 	sandboxv1alpha1 "sigs.k8s.io/agent-sandbox/api/v1alpha1"
@@ -43,16 +45,27 @@ import (
 const (
 	sandboxLabel                = "agents.x-k8s.io/sandbox-name-hash"
 	sandboxControllerFieldOwner = "sandbox-controller"
+	readinessObserved           = "agents.x-k8s.io/readiness-observed"
 )
 
 var (
 	// Scheme for use by sandbox controllers. Registers required types for client.
 	Scheme = runtime.NewScheme()
+
+	sandboxCreationLatency = prometheus.NewHistogram(
+		prometheus.HistogramOpts{
+			Name: "sandbox_creation_latency",
+			Help: "Time taken from sandbox creation to sandbox ready",
+			// 1, 2, 3, 5, 7, 10, 15, 20, 30, 45, 60, 90, 120, 180, 300
+			Buckets: []float64{1, 2, 3, 5, 7, 10, 15, 20, 30, 45, 60, 90, 120, 180, 300},
+		},
+	)
 )
 
 func init() {
 	utilruntime.Must(clientgoscheme.AddToScheme(Scheme))
 	utilruntime.Must(sandboxv1alpha1.AddToScheme(Scheme))
+	metrics.Registry.MustRegister(sandboxCreationLatency)
 }
 
 // SandboxReconciler reconciles a Sandbox object
@@ -114,13 +127,48 @@ func (r *SandboxReconciler) Reconcile(ctx context.Context, req ctrl.Request) (ct
 	// Update status
 	if statusUpdateErr := r.updateStatus(ctx, oldStatus, sandbox); statusUpdateErr != nil {
 		// Surface update error
+		statusUpdateErr = fmt.Errorf("faild to update status: %w", statusUpdateErr)
 		err = errors.Join(err, statusUpdateErr)
 	}
 
+	if recordErr := r.recordFirstReadyMetric(ctx, sandbox); recordErr != nil {
+		err = errors.Join(err, recordErr)
+	}
+
 	// return errors seen
 	return ctrl.Result{RequeueAfter: requeueAfter}, err
 }
 
+func (r *SandboxReconciler) recordFirstReadyMetric(ctx context.Context, sandbox *sandboxv1alpha1.Sandbox) error {
+	log := log.FromContext(ctx)
+
+	// If already recorded ready dont record metric
+	if sandbox.Annotations != nil && sandbox.Annotations[readinessObserved] != "" {
+		return nil
+	}
+
+	// If not ready dont record metric
+	if !meta.IsStatusConditionTrue(sandbox.Status.Conditions, string(sandboxv1alpha1.SandboxConditionReady)) {
+		return nil
+	}
+
+	// record metric
+	latency := time.Since(sandbox.CreationTimestamp.Time).Seconds()
+	sandboxCreationLatency.Observe(latency)
+
+	// add annotation
+	patch := client.MergeFrom(sandbox.DeepCopy())
+	if sandbox.Annotations == nil {
+		sandbox.Annotations = make(map[string]string)
+	}
+	sandbox.Annotations[readinessObserved] = "true"
+	if err := r.Patch(ctx, sandbox, patch); err != nil {
+		log.Error(err, "Failed to add first-ready-metric annotation")
+		return err
+	}
+	return nil
+}
+
 func (r *SandboxReconciler) reconcileChildResources(ctx context.Context, sandbox *sandboxv1alpha1.Sandbox) error {
 	// Create a hash from the sandbox.Name and use it as label value
 	nameHash := NameHash(sandbox.Name)

controllers/sandbox_controller_test.go

@@ -16,14 +16,17 @@ package controllers
 
 import (
 	"errors"
+	"strings"
 	"testing"
 	"time"
 
 	"github.com/google/go-cmp/cmp"
 	"github.com/google/go-cmp/cmp/cmpopts"
+	"github.com/prometheus/client_golang/prometheus/testutil"
 	"github.com/stretchr/testify/require"
 	corev1 "k8s.io/api/core/v1"
 	k8serrors "k8s.io/apimachinery/pkg/api/errors"
+	"k8s.io/apimachinery/pkg/api/meta"
 	"k8s.io/apimachinery/pkg/api/resource"
 	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
 	"k8s.io/apimachinery/pkg/runtime"
@@ -649,3 +652,92 @@ func TestSandboxExpiry(t *testing.T) {
 		})
 	}
 }
+
+func TestSandboxCreationLatencyMetric(t *testing.T) {
+	sandboxName := "sandbox-name"
+	sandboxNs := "sandbox-ns"
+	sb := &sandboxv1alpha1.Sandbox{}
+	sb.Name = sandboxName
+	sb.Namespace = sandboxNs
+	sb.Generation = 1
+	sb.CreationTimestamp = metav1.NewTime(time.Now())
+	sb.Spec = sandboxv1alpha1.SandboxSpec{
+		PodTemplate: sandboxv1alpha1.PodTemplate{
+			Spec: corev1.PodSpec{
+				Containers: []corev1.Container{
+					{
+						Name: "test-container",
+					},
+				},
+			},
+		},
+	}
+
+	r := SandboxReconciler{
+		Client: newFakeClient(sb),
+		Scheme: Scheme,
+	}
+
+	_, err := r.Reconcile(t.Context(), ctrl.Request{
+		NamespacedName: types.NamespacedName{
+			Name:      sandboxName,
+			Namespace: sandboxNs,
+		},
+	})
+	require.NoError(t, err)
+
+	// get pod and mark it ready
+	pod := &corev1.Pod{}
+	require.NoError(t, r.Get(t.Context(), types.NamespacedName{Name: sandboxName, Namespace: sandboxNs}, pod))
+	pod.Status.Phase = corev1.PodRunning
+	pod.Status.Conditions = []corev1.PodCondition{
+		{
+			Type:   corev1.PodReady,
+			Status: corev1.ConditionTrue,
+		},
+	}
+	require.NoError(t, r.Status().Update(t.Context(), pod))
+
+	_, err = r.Reconcile(t.Context(), ctrl.Request{
+		NamespacedName: types.NamespacedName{
+			Name:      sandboxName,
+			Namespace: sandboxNs,
+		},
+	})
+	require.NoError(t, err)
+
+	// Validate Sandbox status
+	liveSandbox := &sandboxv1alpha1.Sandbox{}
+	require.NoError(t, r.Get(t.Context(), types.NamespacedName{Name: sandboxName, Namespace: sandboxNs}, liveSandbox))
+	require.True(t, meta.IsStatusConditionTrue(liveSandbox.Status.Conditions, "Ready"))
+	require.NotNil(t, liveSandbox.Annotations)
+	require.Equal(t, "true", liveSandbox.Annotations[readinessObserved])
+
+	// Check metric
+	expected := `
+	# HELP sandbox_creation_latency Time taken from sandbox creation to sandbox ready
+	# TYPE sandbox_creation_latency histogram
+	sandbox_creation_latency_bucket{le="1"} 1
+	sandbox_creation_latency_bucket{le="2"} 1
+	sandbox_creation_latency_bucket{le="3"} 1
+	sandbox_creation_latency_bucket{le="5"} 1
+	sandbox_creation_latency_bucket{le="7"} 1
+	sandbox_creation_latency_bucket{le="10"} 1
+	sandbox_creation_latency_bucket{le="15"} 1
+	sandbox_creation_latency_bucket{le="20"} 1
+	sandbox_creation_latency_bucket{le="30"} 1
+	sandbox_creation_latency_bucket{le="45"} 1
+	sandbox_creation_latency_bucket{le="60"} 1
+	sandbox_creation_latency_bucket{le="90"} 1
+	sandbox_creation_latency_bucket{le="120"} 1
+	sandbox_creation_latency_bucket{le="180"} 1
+	sandbox_creation_latency_bucket{le="300"} 1
+	sandbox_creation_latency_bucket{le="+Inf"} 1
+	sandbox_creation_latency_count 1
+	`
+	err = testutil.CollectAndCompare(sandboxCreationLatency, strings.NewReader(expected), "sandbox_creation_latency")
+	// We ignore the error because the sum is not deterministic
+	if err != nil && !strings.Contains(err.Error(), "sandbox_creation_latency_sum") {
+		require.NoError(t, err)
+	}
+}

go.mod

@@ -4,6 +4,7 @@ go 1.24.4
 
 require (
 	github.com/google/go-cmp v0.7.0
+	github.com/prometheus/client_golang v1.23.2
 	github.com/stretchr/testify v1.11.1
 	k8s.io/api v0.34.1
 	k8s.io/apiextensions-apiserver v0.34.1
@@ -43,13 +44,13 @@ require (
 	github.com/google/gnostic-models v0.7.0 // indirect
 	github.com/google/uuid v1.6.0 // indirect
 	github.com/json-iterator/go v1.1.12 // indirect
+	github.com/kylelemons/godebug v1.1.0 // indirect
 	github.com/modern-go/concurrent v0.0.0-20180306012644-bacd9c7ef1dd // indirect
 	github.com/modern-go/reflect2 v1.0.3-0.20250322232337-35a7c28c31ee // indirect
 	github.com/munnerz/goautoneg v0.0.0-20191010083416-a7dc8b61c822 // indirect
 	github.com/onsi/ginkgo/v2 v2.23.3 // indirect
 	github.com/onsi/gomega v1.37.0 // indirect
 	github.com/pmezard/go-difflib v1.0.0 // indirect
-	github.com/prometheus/client_golang v1.23.2 // indirect
 	github.com/prometheus/client_model v0.6.2 // indirect
 	github.com/prometheus/common v0.67.1 // indirect
 	github.com/prometheus/procfs v0.17.0 // indirect