constellation/operators/constellation-node-operator/sgreconciler/scalinggroup_controller.go
Malte Poll 4283601433
operators: infrastructure autodiscovery (#1958)
* helm: configure GCP cloud controller manager to search in all zones of a region

See also: d716fdd452/providers/gce/gce.go (L376-L380)

* operators: add nodeGroupName to ScalingGroup CRD

NodeGroupName is the human friendly name of the node group that will be exposed to customers via the Constellation config in the future.

* operators: support simple executor / scheduler to reconcile on non-k8s resources

* operators: add new return type for ListScalingGroups to support arbitrary node groups

* operators: ListScalingGroups should return additionally created node groups on AWS

* operators: ListScalingGroups should return additionally created node groups on Azure

* operators: ListScalingGroups should return additionally created node groups on GCP

* operators: ListScalingGroups should return additionally created node groups on unsupported CSPs

* operators: implement external scaling group reconciler

This controller scans the cloud provider infrastructure and changes k8s resources accordingly.
It creates ScaleSet resources when new node groups are created and deletes them if the node groups are removed.

* operators: no longer create scale sets when the operator starts

In the future, scale sets are created dynamically.

* operators: watch for node join/leave events using a controller

* operators: deploy new controllers

* docs: update auto scaling documentation with support for node groups
2023-07-05 07:27:34 +02:00

186 lines
6.5 KiB
Go

/*
Copyright (c) Edgeless Systems GmbH
SPDX-License-Identifier: AGPL-3.0-only
*/
package sgreconciler
import (
"context"
"strings"
mainconstants "github.com/edgelesssys/constellation/v2/internal/constants"
updatev1alpha1 "github.com/edgelesssys/constellation/v2/operators/constellation-node-operator/v2/api/v1alpha1"
cspapi "github.com/edgelesssys/constellation/v2/operators/constellation-node-operator/v2/internal/cloud/api"
"github.com/edgelesssys/constellation/v2/operators/constellation-node-operator/v2/internal/executor"
k8sErrors "k8s.io/apimachinery/pkg/api/errors"
metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
"k8s.io/client-go/util/retry"
"sigs.k8s.io/controller-runtime/pkg/client"
"sigs.k8s.io/controller-runtime/pkg/log"
)
const (
// defaultScalingGroupMin is the default minimum number of nodes in a scaling group.
// This value is used if the scaling group is created by the operator.
// If a user modifies the scaling group, the operator will not overwrite the user's configuration.
defaultScalingGroupMin = 1
// defaultScalingGroupMax is the default maximum number of nodes in a scaling group.
// This value is used if the scaling group is created by the operator.
// If a user modifies the scaling group, the operator will not overwrite the user's configuration.
defaultScalingGroupMax = 10
)
// ExternalScalingGroupReconciler reconciles on scaling groups in CSP infrastructure.
// It does NOT reconcile on k8s resources.
// Instead, it scans the cloud provider infrastructure and changes k8s resources accordingly.
type ExternalScalingGroupReconciler struct {
// uid is the unique identifier of the Constellation cluster.
uid string
scalingGroupDiscoverer scalingGroupDiscoverer
k8sClient k8sReadWriter
}
// NewExternalScalingGroupReconciler creates a new InfrastructureReconciler.
func NewExternalScalingGroupReconciler(uid string, discoverer scalingGroupDiscoverer, k8sClient k8sReadWriter) *ExternalScalingGroupReconciler {
return &ExternalScalingGroupReconciler{
uid: uid,
scalingGroupDiscoverer: discoverer,
k8sClient: k8sClient,
}
}
// Reconcile reconciles on scaling groups in CSP infrastructure.
func (r *ExternalScalingGroupReconciler) Reconcile(ctx context.Context) (executor.Result, error) {
logr := log.FromContext(ctx)
logr.Info("reconciling external scaling groups")
nodeGroups, err := r.scalingGroupDiscoverer.ListScalingGroups(ctx, r.uid)
if err != nil {
return executor.Result{}, err
}
existingNodeGroups := map[string]struct{}{}
// create all scaling groups that are newly discovered
for _, group := range nodeGroups {
exists, err := patchNodeGroupName(ctx, r.k8sClient, group.Name, group.NodeGroupName)
if err != nil {
return executor.Result{}, err
}
if exists {
// scaling group already exists
existingNodeGroups[group.Name] = struct{}{}
continue
}
err = createScalingGroupIfNotExists(ctx, newScalingGroupConfig{
k8sClient: r.k8sClient,
resourceName: group.Name,
groupID: group.GroupID,
nodeGroupName: group.NodeGroupName,
autoscalingGroupName: group.AutoscalingGroupName,
role: group.Role,
})
if err != nil {
return executor.Result{}, err
}
existingNodeGroups[group.Name] = struct{}{}
}
logr.Info("ensured scaling groups are created", "count", len(nodeGroups))
// delete all scaling groups that no longer exist
var scalingGroups updatev1alpha1.ScalingGroupList
if err := r.k8sClient.List(ctx, &scalingGroups); err != nil {
return executor.Result{}, err
}
for _, group := range scalingGroups.Items {
if _, ok := existingNodeGroups[group.Name]; !ok {
logr.Info("deleting scaling group", "name", group.Name)
err := r.k8sClient.Delete(ctx, &group)
if err != nil {
return executor.Result{}, err
}
}
}
logr.Info("external scaling groups reconciled")
return executor.Result{}, nil
}
// patchNodeGroupName patches the node group name of a scaling group resource (if necessary and it exists).
func patchNodeGroupName(ctx context.Context, k8sClient k8sReadWriter, resourceName, nodeGroupName string) (exists bool, err error) {
logr := log.FromContext(ctx)
var scalingGroup updatev1alpha1.ScalingGroup
err = k8sClient.Get(ctx, client.ObjectKey{Name: resourceName}, &scalingGroup)
if k8sErrors.IsNotFound(err) {
// scaling group does not exist
// no need to patch
return false /* doesn't exist */, nil
}
if err != nil {
return false, err
}
if scalingGroup.Spec.NodeGroupName == nodeGroupName {
// scaling group already has the correct node group name
return true /* exists */, nil
}
logr.Info("patching node group name", "resourceName", resourceName, "nodeGroupName", nodeGroupName)
return true, retry.RetryOnConflict(retry.DefaultRetry, func() error {
if err := k8sClient.Get(ctx, client.ObjectKey{Name: resourceName}, &scalingGroup); err != nil {
return err
}
scalingGroup.Spec.NodeGroupName = nodeGroupName
return k8sClient.Update(ctx, &scalingGroup)
})
}
func createScalingGroupIfNotExists(ctx context.Context, config newScalingGroupConfig) error {
logr := log.FromContext(ctx)
err := config.k8sClient.Create(ctx, &updatev1alpha1.ScalingGroup{
TypeMeta: metav1.TypeMeta{APIVersion: "update.edgeless.systems/v1alpha1", Kind: "ScalingGroup"},
ObjectMeta: metav1.ObjectMeta{
Name: strings.ToLower(config.resourceName),
},
Spec: updatev1alpha1.ScalingGroupSpec{
NodeVersion: mainconstants.NodeVersionResourceName,
GroupID: config.groupID,
AutoscalerGroupName: config.autoscalingGroupName,
NodeGroupName: config.nodeGroupName,
Min: defaultScalingGroupMin,
Max: defaultScalingGroupMax,
Role: config.role,
},
})
if k8sErrors.IsAlreadyExists(err) {
return nil
} else if err == nil {
logr.Info("created scaling group", "name", config.resourceName, "nodeGroupName", config.nodeGroupName)
} else {
logr.Error(err, "failed to create scaling group", "name", config.resourceName, "nodeGroupName", config.nodeGroupName)
}
return err
}
// scalingGroupDiscoverer is used to discover scaling groups in the cloud provider infrastructure.
type scalingGroupDiscoverer interface {
ListScalingGroups(ctx context.Context, uid string,
) ([]cspapi.ScalingGroup, error)
}
type k8sReadWriter interface {
client.Reader
client.Writer
}
type newScalingGroupConfig struct {
k8sClient client.Writer
resourceName string
groupID string
nodeGroupName string
autoscalingGroupName string
role updatev1alpha1.NodeRole
}