actions-runner-controller/controllers/runnerdeployment_controller.go

/*
Copyright 2020 The actions-runner-controller authors.

Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at

    http://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
*/

package controllers

import (
	"context"
	"fmt"
	"hash/fnv"
	"reflect"
	"sort"
	"time"

	"k8s.io/apimachinery/pkg/types"

	"github.com/davecgh/go-spew/spew"
	"github.com/go-logr/logr"
	"k8s.io/apimachinery/pkg/runtime"
	"k8s.io/apimachinery/pkg/util/rand"
	"k8s.io/client-go/tools/record"
	ctrl "sigs.k8s.io/controller-runtime"
	"sigs.k8s.io/controller-runtime/pkg/client"

	corev1 "k8s.io/api/core/v1"
	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"

	"github.com/actions-runner-controller/actions-runner-controller/api/v1alpha1"
	"github.com/actions-runner-controller/actions-runner-controller/controllers/metrics"
)

const (
	LabelKeyRunnerTemplateHash   = "runner-template-hash"
	LabelKeyRunnerDeploymentName = "runner-deployment-name"

	runnerSetOwnerKey = ".metadata.controller"
)

// RunnerDeploymentReconciler reconciles a Runner object
type RunnerDeploymentReconciler struct {
	client.Client
	Log                logr.Logger
	Recorder           record.EventRecorder
	Scheme             *runtime.Scheme
	CommonRunnerLabels []string
	Name               string
}

// +kubebuilder:rbac:groups=actions.summerwind.dev,resources=runnerdeployments,verbs=get;list;watch;create;update;patch;delete
// +kubebuilder:rbac:groups=actions.summerwind.dev,resources=runnerdeployments/finalizers,verbs=get;list;watch;create;update;patch;delete
// +kubebuilder:rbac:groups=actions.summerwind.dev,resources=runnerdeployments/status,verbs=get;update;patch
// +kubebuilder:rbac:groups=actions.summerwind.dev,resources=runnerreplicasets,verbs=get;list;watch;create;update;patch;delete
// +kubebuilder:rbac:groups=actions.summerwind.dev,resources=runnerreplicasets/status,verbs=get;update;patch
// +kubebuilder:rbac:groups=core,resources=events,verbs=create;patch

func (r *RunnerDeploymentReconciler) Reconcile(ctx context.Context, req ctrl.Request) (ctrl.Result, error) {
	log := r.Log.WithValues("runnerdeployment", req.NamespacedName)

	var rd v1alpha1.RunnerDeployment
	if err := r.Get(ctx, req.NamespacedName, &rd); err != nil {
		return ctrl.Result{}, client.IgnoreNotFound(err)
	}

	if !rd.ObjectMeta.DeletionTimestamp.IsZero() {
		return ctrl.Result{}, nil
	}

	metrics.SetRunnerDeployment(rd)

	var myRunnerReplicaSetList v1alpha1.RunnerReplicaSetList
	if err := r.List(ctx, &myRunnerReplicaSetList, client.InNamespace(req.Namespace), client.MatchingFields{runnerSetOwnerKey: req.Name}); err != nil {
		return ctrl.Result{}, err
	}

	myRunnerReplicaSets := myRunnerReplicaSetList.Items

	sort.Slice(myRunnerReplicaSets, func(i, j int) bool {
		return myRunnerReplicaSets[i].GetCreationTimestamp().After(myRunnerReplicaSets[j].GetCreationTimestamp().Time)
	})

	var newestSet *v1alpha1.RunnerReplicaSet

	var oldSets []v1alpha1.RunnerReplicaSet

	if len(myRunnerReplicaSets) > 0 {
		newestSet = &myRunnerReplicaSets[0]
	}

	if len(myRunnerReplicaSets) > 1 {
		oldSets = myRunnerReplicaSets[1:]
	}

	desiredRS, err := r.newRunnerReplicaSet(rd)
	if err != nil {
		r.Recorder.Event(&rd, corev1.EventTypeNormal, "RunnerAutoscalingFailure", err.Error())

		log.Error(err, "Could not create runnerreplicaset")

		return ctrl.Result{}, err
	}

	if newestSet == nil {
		if err := r.Client.Create(ctx, desiredRS); err != nil {
			log.Error(err, "Failed to create runnerreplicaset resource")

			return ctrl.Result{}, err
		}

		log.Info("Created runnerreplicaset", "runnerreplicaset", desiredRS.Name)

		return ctrl.Result{}, nil
	}

	newestTemplateHash, ok := getTemplateHash(newestSet)
	if !ok {
		log.Info("Failed to get template hash of newest runnerreplicaset resource. It must be in an invalid state. Please manually delete the runnerreplicaset so that it is recreated")

		return ctrl.Result{}, nil
	}

	desiredTemplateHash, ok := getTemplateHash(desiredRS)
	if !ok {
		log.Info("Failed to get template hash of desired runnerreplicaset resource. It must be in an invalid state. Please manually delete the runnerreplicaset so that it is recreated")

		return ctrl.Result{}, nil
	}

	if newestTemplateHash != desiredTemplateHash {
		if err := r.Client.Create(ctx, desiredRS); err != nil {
			log.Error(err, "Failed to create runnerreplicaset resource")

			return ctrl.Result{}, err
		}

		log.Info("Created runnerreplicaset", "runnerreplicaset", desiredRS.Name)

		// We requeue in order to clean up old runner replica sets later.
		// Otherwise, they aren't cleaned up until the next re-sync interval.
		return ctrl.Result{RequeueAfter: 5 * time.Second}, nil
	}

	if !reflect.DeepEqual(newestSet.Spec.Selector, desiredRS.Spec.Selector) {
		updateSet := newestSet.DeepCopy()
		updateSet.Spec = *desiredRS.Spec.DeepCopy()

		// A selector update change doesn't trigger replicaset replacement,
		// but we still need to update the existing replicaset with it.
		// Otherwise selector-based runner query will never work on replicasets created before the controller v0.17.0
		// See https://github.com/actions-runner-controller/actions-runner-controller/pull/355#discussion_r585379259
		if err := r.Client.Update(ctx, updateSet); err != nil {
			log.Error(err, "Failed to update runnerreplicaset resource")

			return ctrl.Result{}, err
		}

		log.V(1).Info("Updated runnerreplicaset due to selector change")

		// At this point, we are already sure that there's no need to create a new replicaset
		// as the runner template hash is not changed.
		//
		// But we still need to requeue for the (possibly rare) cases that there are still old replicasets that needs
		// to be cleaned up.
		return ctrl.Result{RequeueAfter: 5 * time.Second}, nil
	}

	const defaultReplicas = 1

	currentDesiredReplicas := getIntOrDefault(newestSet.Spec.Replicas, defaultReplicas)
	newDesiredReplicas := getIntOrDefault(desiredRS.Spec.Replicas, defaultReplicas)

	// Please add more conditions that we can in-place update the newest runnerreplicaset without disruption
	//
	// If we missed taking the EffectiveTime diff into account, you might end up experiencing scale-ups being delayed scale-down.
	// See https://github.com/actions-runner-controller/actions-runner-controller/pull/1477#issuecomment-1164154496
	var et1, et2 time.Time
	if newestSet.Spec.EffectiveTime != nil {
		et1 = newestSet.Spec.EffectiveTime.Time
	}
	if rd.Spec.EffectiveTime != nil {
		et2 = rd.Spec.EffectiveTime.Time
	}
	if currentDesiredReplicas != newDesiredReplicas || et1 != et2 {
		newestSet.Spec.Replicas = &newDesiredReplicas
		newestSet.Spec.EffectiveTime = rd.Spec.EffectiveTime

		if err := r.Client.Update(ctx, newestSet); err != nil {
			log.Error(err, "Failed to update runnerreplicaset resource")

			return ctrl.Result{}, err
		}

		log.V(1).Info("Updated runnerreplicaset due to spec change",
			"currentDesiredReplicas", currentDesiredReplicas,
			"newDesiredReplicas", newDesiredReplicas,
			"currentEffectiveTime", newestSet.Spec.EffectiveTime,
			"newEffectiveTime", rd.Spec.EffectiveTime,
		)

		return ctrl.Result{}, err
	}

	// Do we have old runner replica sets that should eventually deleted?
	if len(oldSets) > 0 {
		var readyReplicas int
		if newestSet.Status.ReadyReplicas != nil {
			readyReplicas = *newestSet.Status.ReadyReplicas
		}

		oldSetsCount := len(oldSets)

		logWithDebugInfo := log.WithValues(
			"newest_runnerreplicaset", types.NamespacedName{
				Namespace: newestSet.Namespace,
				Name:      newestSet.Name,
			},
			"newest_runnerreplicaset_replicas_ready", readyReplicas,
			"newest_runnerreplicaset_replicas_desired", currentDesiredReplicas,
			"old_runnerreplicasets_count", oldSetsCount,
		)

		if readyReplicas < currentDesiredReplicas {
			logWithDebugInfo.
				Info("Waiting until the newest runnerreplicaset to be 100% available")

			return ctrl.Result{}, nil
		}

		if oldSetsCount > 0 {
			logWithDebugInfo.
				Info("The newest runnerreplicaset is 100% available. Deleting old runnerreplicasets")
		}

		for i := range oldSets {
			rs := oldSets[i]

			rslog := log.WithValues("runnerreplicaset", rs.Name)

			if rs.Status.Replicas != nil && *rs.Status.Replicas > 0 {
				if rs.Spec.Replicas != nil && *rs.Spec.Replicas == 0 {
					rslog.V(2).Info("Waiting for runnerreplicaset to scale to zero")

					continue
				}

				updated := rs.DeepCopy()
				zero := 0
				updated.Spec.Replicas = &zero
				if err := r.Client.Update(ctx, updated); err != nil {
					rslog.Error(err, "Failed to scale runnerreplicaset to zero")

					return ctrl.Result{}, err
				}

				rslog.Info("Scaled runnerreplicaset to zero")

				continue
			}

			if err := r.Client.Delete(ctx, &rs); err != nil {
				rslog.Error(err, "Failed to delete runnerreplicaset resource")

				return ctrl.Result{}, err
			}

			r.Recorder.Event(&rd, corev1.EventTypeNormal, "RunnerReplicaSetDeleted", fmt.Sprintf("Deleted runnerreplicaset '%s'", rs.Name))

			rslog.Info("Deleted runnerreplicaset")
		}
	}

	var replicaSets []v1alpha1.RunnerReplicaSet

	replicaSets = append(replicaSets, *newestSet)
	replicaSets = append(replicaSets, oldSets...)

	var totalCurrentReplicas, totalStatusAvailableReplicas, updatedReplicas int

	for _, rs := range replicaSets {
		var current, available int

		if rs.Status.Replicas != nil {
			current = *rs.Status.Replicas
		}

		if rs.Status.AvailableReplicas != nil {
			available = *rs.Status.AvailableReplicas
		}

		totalCurrentReplicas += current
		totalStatusAvailableReplicas += available
	}

	if newestSet.Status.Replicas != nil {
		updatedReplicas = *newestSet.Status.Replicas
	}

	var status v1alpha1.RunnerDeploymentStatus

	status.AvailableReplicas = &totalStatusAvailableReplicas
	status.ReadyReplicas = &totalStatusAvailableReplicas
	status.DesiredReplicas = &newDesiredReplicas
	status.Replicas = &totalCurrentReplicas
	status.UpdatedReplicas = &updatedReplicas

	if !reflect.DeepEqual(rd.Status, status) {
		updated := rd.DeepCopy()
		updated.Status = status

		if err := r.Status().Patch(ctx, updated, client.MergeFrom(&rd)); err != nil {
			log.Info("Failed to patch runnerdeployment status. Retrying immediately", "error", err.Error())
			return ctrl.Result{
				Requeue: true,
			}, nil
		}
	}

	return ctrl.Result{}, nil
}

func getIntOrDefault(p *int, d int) int {
	if p == nil {
		return d
	}

	return *p
}

func getTemplateHash(rs *v1alpha1.RunnerReplicaSet) (string, bool) {
	hash, ok := rs.Labels[LabelKeyRunnerTemplateHash]

	return hash, ok
}

// ComputeHash returns a hash value calculated from pod template and
// a collisionCount to avoid hash collision. The hash will be safe encoded to
// avoid bad words.
//
// Proudly modified and adopted from k8s.io/kubernetes/pkg/util/hash.DeepHashObject and
// k8s.io/kubernetes/pkg/controller.ComputeHash.
func ComputeHash(template interface{}) string {
	hasher := fnv.New32a()

	hasher.Reset()

	printer := spew.ConfigState{
		Indent:         " ",
		SortKeys:       true,
		DisableMethods: true,
		SpewKeys:       true,
	}
	printer.Fprintf(hasher, "%#v", template)

	return rand.SafeEncodeString(fmt.Sprint(hasher.Sum32()))
}

// Clones the given map and returns a new map with the given key and value added.
// Returns the given map, if labelKey is empty.
//
// Proudly copied from k8s.io/kubernetes/pkg/util/labels.CloneAndAddLabel
func CloneAndAddLabel(labels map[string]string, labelKey, labelValue string) map[string]string {
	if labelKey == "" {
		// Don't need to add a label.
		return labels
	}
	// Clone.
	newLabels := map[string]string{}
	for key, value := range labels {
		newLabels[key] = value
	}
	newLabels[labelKey] = labelValue
	return newLabels
}

// Clones the given selector and returns a new selector with the given key and value added.
// Returns the given selector, if labelKey is empty.
//
// Proudly copied from k8s.io/kubernetes/pkg/util/labels.CloneSelectorAndAddLabel
func CloneSelectorAndAddLabel(selector *metav1.LabelSelector, labelKey, labelValue string) *metav1.LabelSelector {
	if labelKey == "" {
		// Don't need to add a label.
		return selector
	}

	// Clone.
	newSelector := new(metav1.LabelSelector)

	newSelector.MatchLabels = make(map[string]string)
	if selector.MatchLabels != nil {
		for key, val := range selector.MatchLabels {
			newSelector.MatchLabels[key] = val
		}
	}
	newSelector.MatchLabels[labelKey] = labelValue

	if selector.MatchExpressions != nil {
		newMExps := make([]metav1.LabelSelectorRequirement, len(selector.MatchExpressions))
		for i, me := range selector.MatchExpressions {
			newMExps[i].Key = me.Key
			newMExps[i].Operator = me.Operator
			if me.Values != nil {
				newMExps[i].Values = make([]string, len(me.Values))
				copy(newMExps[i].Values, me.Values)
			} else {
				newMExps[i].Values = nil
			}
		}
		newSelector.MatchExpressions = newMExps
	} else {
		newSelector.MatchExpressions = nil
	}

	return newSelector
}

func (r *RunnerDeploymentReconciler) newRunnerReplicaSet(rd v1alpha1.RunnerDeployment) (*v1alpha1.RunnerReplicaSet, error) {
	return newRunnerReplicaSet(&rd, r.CommonRunnerLabels, r.Scheme)
}

func getSelector(rd *v1alpha1.RunnerDeployment) *metav1.LabelSelector {
	selector := rd.Spec.Selector
	if selector == nil {
		selector = &metav1.LabelSelector{MatchLabels: map[string]string{LabelKeyRunnerDeploymentName: rd.Name}}
	}

	return selector
}

func newRunnerReplicaSet(rd *v1alpha1.RunnerDeployment, commonRunnerLabels []string, scheme *runtime.Scheme) (*v1alpha1.RunnerReplicaSet, error) {
	newRSTemplate := *rd.Spec.Template.DeepCopy()

	newRSTemplate.Spec.Labels = append(newRSTemplate.Spec.Labels, commonRunnerLabels...)

	templateHash := ComputeHash(&newRSTemplate)

	// Add template hash label to selector.
	newRSTemplate.ObjectMeta.Labels = CloneAndAddLabel(newRSTemplate.ObjectMeta.Labels, LabelKeyRunnerTemplateHash, templateHash)

	// This label selector is used by default when rd.Spec.Selector is empty.
	newRSTemplate.ObjectMeta.Labels = CloneAndAddLabel(newRSTemplate.ObjectMeta.Labels, LabelKeyRunnerDeploymentName, rd.Name)

	selector := getSelector(rd)

	newRSSelector := CloneSelectorAndAddLabel(selector, LabelKeyRunnerTemplateHash, templateHash)

	rs := v1alpha1.RunnerReplicaSet{
		TypeMeta: metav1.TypeMeta{},
		ObjectMeta: metav1.ObjectMeta{
			GenerateName: rd.ObjectMeta.Name + "-",
			Namespace:    rd.ObjectMeta.Namespace,
			Labels:       newRSTemplate.ObjectMeta.Labels,
		},
		Spec: v1alpha1.RunnerReplicaSetSpec{
			Replicas:      rd.Spec.Replicas,
			Selector:      newRSSelector,
			Template:      newRSTemplate,
			EffectiveTime: rd.Spec.EffectiveTime,
		},
	}

	if err := ctrl.SetControllerReference(rd, &rs, scheme); err != nil {
		return &rs, err
	}

	return &rs, nil
}

func (r *RunnerDeploymentReconciler) SetupWithManager(mgr ctrl.Manager) error {
	name := "runnerdeployment-controller"
	if r.Name != "" {
		name = r.Name
	}

	r.Recorder = mgr.GetEventRecorderFor(name)

	if err := mgr.GetFieldIndexer().IndexField(context.TODO(), &v1alpha1.RunnerReplicaSet{}, runnerSetOwnerKey, func(rawObj client.Object) []string {
		runnerSet := rawObj.(*v1alpha1.RunnerReplicaSet)
		owner := metav1.GetControllerOf(runnerSet)
		if owner == nil {
			return nil
		}

		if owner.APIVersion != v1alpha1.GroupVersion.String() || owner.Kind != "RunnerDeployment" {
			return nil
		}

		return []string{owner.Name}
	}); err != nil {
		return err
	}

	return ctrl.NewControllerManagedBy(mgr).
		For(&v1alpha1.RunnerDeployment{}).
		Owns(&v1alpha1.RunnerReplicaSet{}).
		Named(name).
		Complete(r)
}