Tolerations support in otelcol CRD (open-telemetry#302)

Signed-off-by: Vineeth Pothulapati <[email protected]>

api/v1alpha1/opentelemetrycollector_types.go

@@ -87,6 +87,12 @@ type OpenTelemetryCollectorSpec struct {
 	// +optional
 	// +operator-sdk:gen-csv:customresourcedefinitions.specDescriptors=true
 	Resources v1.ResourceRequirements `json:"resources,omitempty"`
+
+	// Toleration to schedule OpenTelemetry Collector pods.
+	// This is only relevant to daemonsets, statefulsets and deployments
+	// +optional
+	// +operator-sdk:gen-csv:customresourcedefinitions.specDescriptors=true
+	Tolerations []v1.Toleration `json:"tolerations,omitempty"`
 }
 
 // OpenTelemetryCollectorStatus defines the observed state of OpenTelemetryCollector.

api/v1alpha1/opentelemetrycollector_webhook.go

@@ -15,7 +15,7 @@
 package v1alpha1
 
 import (
-	"errors"
+	"fmt"
 
 	"k8s.io/apimachinery/pkg/runtime"
 	ctrl "sigs.k8s.io/controller-runtime"
@@ -59,23 +59,36 @@ var _ webhook.Validator = &OpenTelemetryCollector{}
 // ValidateCreate implements webhook.Validator so a webhook will be registered for the type.
 func (r *OpenTelemetryCollector) ValidateCreate() error {
 	opentelemetrycollectorlog.Info("validate create", "name", r.Name)
-	if r.Spec.Mode != ModeStatefulSet && len(r.Spec.VolumeClaimTemplates) > 0 {
-		return errors.New("Can only specify VolumeClaimTemplates for statefulsets")
-	}
-	return nil
+	return r.validateCRDSpec()
 }
 
 // ValidateUpdate implements webhook.Validator so a webhook will be registered for the type.
 func (r *OpenTelemetryCollector) ValidateUpdate(old runtime.Object) error {
 	opentelemetrycollectorlog.Info("validate update", "name", r.Name)
-	if r.Spec.Mode != ModeStatefulSet && len(r.Spec.VolumeClaimTemplates) > 0 {
-		return errors.New("Can only specify VolumeClaimTemplates for statefulsets")
-	}
-	return nil
+	return r.validateCRDSpec()
 }
 
 // ValidateDelete implements webhook.Validator so a webhook will be registered for the type.
 func (r *OpenTelemetryCollector) ValidateDelete() error {
 	opentelemetrycollectorlog.Info("validate delete", "name", r.Name)
 	return nil
 }
+
+func (r *OpenTelemetryCollector) validateCRDSpec() error {
+	// validate volumeClaimTemplates
+	if r.Spec.Mode != ModeStatefulSet && len(r.Spec.VolumeClaimTemplates) > 0 {
+		return fmt.Errorf("the OpenTelemetry Collector mode is set to %s, which does not support the attribute 'volumeClaimTemplates'", r.Spec.Mode)
+	}
+
+	// validate replicas
+	if (r.Spec.Mode == ModeSidecar || r.Spec.Mode == ModeDaemonSet) && r.Spec.Replicas != nil {
+		return fmt.Errorf("the OpenTelemetry Collector mode is set to %s, which does not support the attribute 'replicas'", r.Spec.Mode)
+	}
+
+	// validate tolerations
+	if r.Spec.Mode == ModeSidecar && len(r.Spec.Tolerations) > 0 {
+		return fmt.Errorf("the OpenTelemetry Collector mode is set to %s, which does not support the attribute 'tolerations'", r.Spec.Mode)
+	}
+
+	return nil
+}

config/crd/bases/opentelemetry.io_opentelemetrycollectors.yaml

@@ -280,6 +280,46 @@ spec:
                 description: ServiceAccount indicates the name of an existing service
                   account to use with this instance.
                 type: string
+              tolerations:
+                description: Toleration to schedule OpenTelemetry Collector pods.
+                items:
+                  description: The pod this Toleration is attached to tolerates any
+                    taint that matches the triple <key,value,effect> using the matching
+                    operator <operator>.
+                  properties:
+                    effect:
+                      description: Effect indicates the taint effect to match. Empty
+                        means match all taint effects. When specified, allowed values
+                        are NoSchedule, PreferNoSchedule and NoExecute.
+                      type: string
+                    key:
+                      description: Key is the taint key that the toleration applies
+                        to. Empty means match all taint keys. If the key is empty,
+                        operator must be Exists; this combination means to match all
+                        values and all keys.
+                      type: string
+                    operator:
+                      description: Operator represents a key's relationship to the
+                        value. Valid operators are Exists and Equal. Defaults to Equal.
+                        Exists is equivalent to wildcard for value, so that a pod
+                        can tolerate all taints of a particular category.
+                      type: string
+                    tolerationSeconds:
+                      description: TolerationSeconds represents the period of time
+                        the toleration (which must be of effect NoExecute, otherwise
+                        this field is ignored) tolerates the taint. By default, it
+                        is not set, which means tolerate the taint forever (do not
+                        evict). Zero and negative values will be treated as 0 (evict
+                        immediately) by the system.
+                      format: int64
+                      type: integer
+                    value:
+                      description: Value is the taint value the toleration matches
+                        to. If the operator is Exists, the value should be empty,
+                        otherwise just a regular string.
+                      type: string
+                  type: object
+                type: array
               volumeClaimTemplates:
                 description: VolumeClaimTemplates will provide stable storage using
                   PersistentVolumes. Only available when the mode=statefulset.

pkg/collector/daemonset.go

@@ -52,6 +52,7 @@ func DaemonSet(cfg config.Config, logger logr.Logger, otelcol v1alpha1.OpenTelem
 					ServiceAccountName: ServiceAccountName(otelcol),
 					Containers:         []corev1.Container{Container(cfg, logger, otelcol)},
 					Volumes:            Volumes(cfg, otelcol),
+					Tolerations:        otelcol.Spec.Tolerations,
 				},
 			},
 		},

pkg/collector/daemonset_test.go

@@ -31,6 +31,9 @@ func TestDaemonSetNewDefault(t *testing.T) {
 		ObjectMeta: metav1.ObjectMeta{
 			Name: "my-instance",
 		},
+		Spec: v1alpha1.OpenTelemetryCollectorSpec{
+			Tolerations: testTolerationValues,
+		},
 	}
 	cfg := config.New()
 
@@ -43,6 +46,7 @@ func TestDaemonSetNewDefault(t *testing.T) {
 	assert.Equal(t, "true", d.Annotations["prometheus.io/scrape"])
 	assert.Equal(t, "8888", d.Annotations["prometheus.io/port"])
 	assert.Equal(t, "/metrics", d.Annotations["prometheus.io/path"])
+	assert.Equal(t, testTolerationValues, d.Spec.Template.Spec.Tolerations)
 
 	assert.Len(t, d.Spec.Template.Spec.Containers, 1)
 

pkg/collector/deployment.go

@@ -53,6 +53,7 @@ func Deployment(cfg config.Config, logger logr.Logger, otelcol v1alpha1.OpenTele
 					ServiceAccountName: ServiceAccountName(otelcol),
 					Containers:         []corev1.Container{Container(cfg, logger, otelcol)},
 					Volumes:            Volumes(cfg, otelcol),
+					Tolerations:        otelcol.Spec.Tolerations,
 				},
 			},
 		},

pkg/collector/deployment_test.go

@@ -17,21 +17,33 @@ package collector_test
 import (
 	"testing"
 
-	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
-
 	"github.com/stretchr/testify/assert"
 
+	v1 "k8s.io/api/core/v1"
+	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
+
 	"github.com/open-telemetry/opentelemetry-operator/api/v1alpha1"
 	"github.com/open-telemetry/opentelemetry-operator/internal/config"
 	. "github.com/open-telemetry/opentelemetry-operator/pkg/collector"
 )
 
+var testTolerationValues = []v1.Toleration{
+	{
+		Key:    "hii",
+		Value:  "greeting",
+		Effect: "NoSchedule",
+	},
+}
+
 func TestDeploymentNewDefault(t *testing.T) {
 	// prepare
 	otelcol := v1alpha1.OpenTelemetryCollector{
 		ObjectMeta: metav1.ObjectMeta{
 			Name: "my-instance",
 		},
+		Spec: v1alpha1.OpenTelemetryCollectorSpec{
+			Tolerations: testTolerationValues,
+		},
 	}
 	cfg := config.New()
 
@@ -44,6 +56,7 @@ func TestDeploymentNewDefault(t *testing.T) {
 	assert.Equal(t, "true", d.Annotations["prometheus.io/scrape"])
 	assert.Equal(t, "8888", d.Annotations["prometheus.io/port"])
 	assert.Equal(t, "/metrics", d.Annotations["prometheus.io/path"])
+	assert.Equal(t, testTolerationValues, d.Spec.Template.Spec.Tolerations)
 
 	assert.Len(t, d.Spec.Template.Spec.Containers, 1)
 

pkg/collector/statefulset.go

@@ -53,6 +53,7 @@ func StatefulSet(cfg config.Config, logger logr.Logger, otelcol v1alpha1.OpenTel
 					ServiceAccountName: ServiceAccountName(otelcol),
 					Containers:         []corev1.Container{Container(cfg, logger, otelcol)},
 					Volumes:            Volumes(cfg, otelcol),
+					Tolerations:        otelcol.Spec.Tolerations,
 				},
 			},
 			Replicas:             otelcol.Spec.Replicas,

pkg/collector/statefulset_test.go

@@ -35,7 +35,8 @@ func TestStatefulSetNewDefault(t *testing.T) {
 			Name: "my-instance",
 		},
 		Spec: v1alpha1.OpenTelemetryCollectorSpec{
-			Mode: "statefulset",
+			Mode:        "statefulset",
+			Tolerations: testTolerationValues,
 		},
 	}
 	cfg := config.New()
@@ -49,6 +50,7 @@ func TestStatefulSetNewDefault(t *testing.T) {
 	assert.Equal(t, "true", ss.Annotations["prometheus.io/scrape"])
 	assert.Equal(t, "8888", ss.Annotations["prometheus.io/port"])
 	assert.Equal(t, "/metrics", ss.Annotations["prometheus.io/path"])
+	assert.Equal(t, testTolerationValues, ss.Spec.Template.Spec.Tolerations)
 
 	assert.Len(t, ss.Spec.Template.Spec.Containers, 1)