🌱 Release v3.2.0

docs/book/src/SUMMARY.md

@@ -73,6 +73,9 @@
 
   - [Generating CRDs](./reference/generating-crd.md)
   - [Using Finalizers](./reference/using-finalizers.md)
+  - [Watching Resources](./reference/watching-resources.md)
+    - [Resources Managed by the Operator](./reference/watching-resources/operator-managed.md)
+    - [Externally Managed Resources](./reference/watching-resources/externally-managed.md)
   - [Kind cluster](reference/kind.md)
   - [What's a webhook?](reference/webhook-overview.md)
     - [Admission webhook](reference/admission-webhook.md)

docs/book/src/reference/reference.md

@@ -5,6 +5,11 @@
     Finalizers are a mechanism to
     execute any custom logic related to a resource before it gets deleted from
     Kubernetes cluster.
+  - [Watching Resources](watching-resources.md)
+    Watch resources in the Kubernetes cluster to be informed and take actions on changes.
+      - [Resources Managed by the Operator](watching-resources/operator-managed.md)
+      - [Externally Managed Resources](watching-resources/externally-managed.md)
+        Controller Runtime provides the ability to watch additional resources relevant to the controlled ones.
   - [Kind cluster](kind.md)
   - [What's a webhook?](webhook-overview.md)
     Webhooks are HTTP callbacks, there are 3

docs/book/src/reference/watching-resources.md

@@ -0,0 +1,16 @@
+# Watching Resources
+
+Inside a `Reconcile()` control loop, you are looking to do a collection of operations until it has the desired state on the cluster.
+Therefore, it can be necessary to know when a resource that you care about is changed.
+In the case that there is an action (create, update, edit, delete, etc.) on a watched resource, `Reconcile()` should be called for the resources watching it.
+
+[Controller Runtime libraries](https://pkg.go.dev/sigs.k8s.io/controller-runtime/pkg/builder) provide many ways for resources to be managed and watched.
+This ranges from the easy and obvious use cases, such as watching the resources which were created and managed by the controller, to more unique and advanced use cases.
+
+See each subsection for explanations and examples of the different ways in which your controller can _Watch_ the resources it cares about.
+
+- [Watching Operator Managed Resources](watching-resources/operator-managed.md) -
+  These resources are created and managed by the same operator as the resource watching them.
+  This section covers both if they are managed by the same controller or separate controllers.
+- [Watching Externally Managed Resources](watching-resources/externally-managed.md) -
+  These resources could be manually created, or managed by other operators/controllers or the Kubernetes control plane.

docs/book/src/reference/watching-resources/externally-managed.md

@@ -0,0 +1,31 @@
+# Watching Externally Managed Resources
+
+By default, Kubebuilder and the Controller Runtime libraries allow for controllers
+to easily watch the resources that they manage as well as dependent resources that are `Owned` by the controller.
+However, those are not always the only resources that need to be watched in the cluster.
+
+## User Specified Resources
+
+There are many examples of Resource Specs that allow users to reference external resources.
+- Ingresses have references to Service objects
+- Pods have references to ConfigMaps, Secrets and Volumes
+- Deployments and Services have references to Pods
+
+This same functionality can be added to CRDs and custom controllers.
+This will allow for resources to be reconciled when another resources it references is changed.
+
+As an example, we are going to create a `ConfigDeployment` resource.
+The `ConfigDeployment`'s purpose is to manage a `Deployment` whose pods are always using the latest version of a `ConfigMap`.
+While ConfigMaps are auto-updated within Pods, applications may not always be able to auto-refresh config from the file system.
+Some applications require restarts to apply configuration updates.
+- The `ConfigDeployment` CRD will hold a reference to a ConfigMap inside its Spec.
+- The `ConfigDeployment` controller will be in charge of creating a deployment with Pods that use the ConfigMap.
+These pods should be updated anytime that the referenced ConfigMap changes, therefore the ConfigDeployments will need to be reconciled on changes to the referenced ConfigMap.
+
+### Allow for linking of resources in the `Spec`
+
+{{#literatego ./testdata/external-indexed-field/api.go}}
+
+### Watch linked resources
+
+{{#literatego ./testdata/external-indexed-field/controller.go}}

docs/book/src/reference/watching-resources/operator-managed.md

@@ -0,0 +1,24 @@
+# Watching Operator Managed Resources
+
+Kubebuilder and the Controller Runtime libraries allow for controllers
+to implement the logic of their CRD through easy management of Kubernetes resources.
+
+## Controlled & Owned Resources
+
+Managing dependency resources is fundamental to a controller, and it's not possible to manage them without watching for changes to their state.
+- Deployments must know when the ReplicaSets that they manage are changed
+- ReplicaSets must know when their Pods are deleted, or change from healthy to unhealthy.
+
+Through the `Owns()` functionality, Controller Runtime provides an easy way to watch dependency resources for changes.
+
+As an example, we are going to create a `SimpleDeployment` resource.
+The `SimpleDeployment`'s purpose is to manage a `Deployment` that users can change certain aspects of, through the `SimpleDeployment` Spec.
+The `SimpleDeployment` controller's purpose is to make sure that it's owned `Deployment` always uses the settings provided by the user.
+
+### Provide basic templating in the `Spec`
+
+{{#literatego ./testdata/owned-resource/api.go}}
+
+### Manage the Owned Resource
+
+{{#literatego ./testdata/owned-resource/controller.go}}

docs/book/src/reference/watching-resources/testdata/external-indexed-field/api.go

@@ -0,0 +1,77 @@
+/*
+
+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.
+*/
+// +kubebuilder:docs-gen:collapse=Apache License
+/* */
+package external_indexed_field
+
+import (
+	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
+)
+// +kubebuilder:docs-gen:collapse=Imports
+
+/*
+In our type's Spec, we want to allow the user to pass in a reference to a configMap in the same namespace.
+It's also possible for this to be a namespaced reference, but in this example we will assume that the referenced object
+lives in the same namespace.
+
+This field does not need to be optional.
+If the field is required, the indexing code in the controller will need to be modified.
+*/
+
+// ConfigDeploymentSpec defines the desired state of ConfigDeployment
+type ConfigDeploymentSpec struct {
+	// INSERT ADDITIONAL SPEC FIELDS - desired state of cluster
+	// Important: Run "make" to regenerate code after modifying this file
+
+	// Name of an existing ConfigMap in the same namespace, to add to the deployment
+	// +optional
+	ConfigMap string `json:"configMap,omitempty"`
+}
+
+/*
+The rest of the API configuration is covered in the CronJob tutorial.
+*/
+/* */
+// ConfigDeploymentStatus defines the observed state of ConfigDeployment
+type ConfigDeploymentStatus struct {
+	// INSERT ADDITIONAL STATUS FIELD - define observed state of cluster
+	// Important: Run "make" to regenerate code after modifying this file
+}
+
+//+kubebuilder:object:root=true
+//+kubebuilder:subresource:status
+
+// ConfigDeployment is the Schema for the configdeployments API
+type ConfigDeployment struct {
+	metav1.TypeMeta   `json:",inline"`
+	metav1.ObjectMeta `json:"metadata,omitempty"`
+
+	Spec   ConfigDeploymentSpec   `json:"spec,omitempty"`
+	Status ConfigDeploymentStatus `json:"status,omitempty"`
+}
+
+//+kubebuilder:object:root=true
+
+// ConfigDeploymentList contains a list of ConfigDeployment
+type ConfigDeploymentList struct {
+	metav1.TypeMeta `json:",inline"`
+	metav1.ListMeta `json:"metadata,omitempty"`
+	Items           []ConfigDeployment `json:"items"`
+}
+
+func init() {
+	SchemeBuilder.Register(&ConfigDeployment{}, &ConfigDeploymentList{})
+}
+// +kubebuilder:docs-gen:collapse=Remaining API Code

docs/book/src/reference/watching-resources/testdata/external-indexed-field/controller.go

@@ -0,0 +1,208 @@
+/*
+
+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.
+*/
+// +kubebuilder:docs-gen:collapse=Apache License
+
+/*
+Along with the standard imports, we need additional controller-runtime and apimachinery libraries.
+All additional libraries, necessary for Watching, have the comment `Required For Watching` appended.
+*/
+package external_indexed_field
+
+import (
+	"context"
+
+	"github.com/go-logr/logr"
+	kapps "k8s.io/api/apps/v1"
+	corev1 "k8s.io/api/core/v1"
+	"k8s.io/apimachinery/pkg/fields" // Required for Watching
+	"k8s.io/apimachinery/pkg/runtime"
+	"k8s.io/apimachinery/pkg/types" // Required for Watching
+	ctrl "sigs.k8s.io/controller-runtime"
+	"sigs.k8s.io/controller-runtime/pkg/builder" // Required for Watching
+	"sigs.k8s.io/controller-runtime/pkg/client"
+	"sigs.k8s.io/controller-runtime/pkg/handler" // Required for Watching
+	"sigs.k8s.io/controller-runtime/pkg/predicate" // Required for Watching
+	"sigs.k8s.io/controller-runtime/pkg/reconcile" // Required for Watching
+	"sigs.k8s.io/controller-runtime/pkg/source" // Required for Watching
+
+	appsv1 "tutorial.kubebuilder.io/project/api/v1"
+)
+
+/*
+Determine the path of the field in the ConfigDeployment CRD that we wish to use as the "object reference".
+This will be used in both the indexing and watching.
+*/
+const (
+	configMapField = ".spec.configMap"
+)
+
+/*
+*/
+
+// ConfigDeploymentReconciler reconciles a ConfigDeployment object
+type ConfigDeploymentReconciler struct {
+	client.Client
+	Log    logr.Logger
+	Scheme *runtime.Scheme
+}
+// +kubebuilder:docs-gen:collapse=Reconciler Declaration
+
+/*
+There are two additional resources that the controller needs to have access to, other than ConfigDeployments.
+- It needs to be able to fully manage Deployments, as well as check their status.
+- It also needs to be able to get, list and watch ConfigMaps.
+All 3 of these are important, and you will see usages of each below.
+*/
+
+//+kubebuilder:rbac:groups=apps.tutorial.kubebuilder.io,resources=configdeployments,verbs=get;list;watch;create;update;patch;delete
+//+kubebuilder:rbac:groups=apps.tutorial.kubebuilder.io,resources=configdeployments/status,verbs=get;update;patch
+//+kubebuilder:rbac:groups=apps.tutorial.kubebuilder.io,resources=configdeployments/finalizers,verbs=update
+//+kubebuilder:rbac:groups=apps,resources=deployments,verbs=get;list;watch;create;update;patch;delete
+//+kubebuilder:rbac:groups=apps,resources=deployments/status,verbs=get
+//+kubebuilder:rbac:groups="",resources=configmaps,verbs=get;list;watch
+
+/*
+`Reconcile` will be in charge of reconciling the state of ConfigDeployments.
+ConfigDeployments are used to manage Deployments whose pods are updated whenever the configMap that they use is updated.
+
+For that reason we need to add an annotation to the PodTemplate within the Deployment we create.
+This annotation will keep track of the latest version of the data within the referenced ConfigMap.
+Therefore when the version of the configMap is changed, the PodTemplate in the Deployment will change.
+This will cause a rolling upgrade of all Pods managed by the Deployment.
+
+Skip down to the `SetupWithManager` function to see how we ensure that `Reconcile` is called when the referenced `ConfigMaps` are updated.
+*/
+// Reconcile is part of the main kubernetes reconciliation loop which aims to
+// move the current state of the cluster closer to the desired state.
+func (r *ConfigDeploymentReconciler) Reconcile(ctx context.Context, req ctrl.Request) (ctrl.Result, error) {
+	/* */
+	log := r.Log.WithValues("configDeployment", req.NamespacedName)
+
+	var configDeployment appsv1.ConfigDeployment
+	if err := r.Get(ctx, req.NamespacedName, &configDeployment); err != nil {
+		log.Error(err, "unable to fetch ConfigDeployment")
+		// we'll ignore not-found errors, since they can't be fixed by an immediate
+		// requeue (we'll need to wait for a new notification), and we can get them
+		// on deleted requests.
+		return ctrl.Result{}, client.IgnoreNotFound(err)
+	}
+	// +kubebuilder:docs-gen:collapse=Begin the Reconcile
+
+	// your logic here
+
+	var configMapVersion string
+	if configDeployment.Spec.ConfigMap != "" {
+		configMapName := configDeployment.Spec.ConfigMap
+		foundConfigMap := &corev1.ConfigMap{}
+		err := r.Get(ctx, types.NamespacedName{Name: configMapName, Namespace: configDeployment.Namespace}, foundConfigMap)
+		if err != nil {
+			// If a configMap name is provided, then it must exist
+			// You will likely want to create an Event for the user to understand why their reconcile is failing.
+			return ctrl.Result{}, err
+		}
+
+		// Hash the data in some way, or just use the version of the Object
+		configMapVersion = foundConfigMap.ResourceVersion
+	}
+
+	// Logic here to add the configMapVersion as an annotation on your Deployment Pods.
+
+	return ctrl.Result{}, nil
+}
+
+/*
+Finally, we add this reconciler to the manager, so that it gets started
+when the manager is started.
+
+Since we create dependency Deployments during the reconcile, we can specify that the controller `Owns` Deployments.
+
+However the ConfigMaps that we want to watch are not owned by the ConfigDeployment object.
+Therefore we must specify a custom way of watching those objects.
+This watch logic is complex, so we have split it into a separate method.
+*/
+
+// SetupWithManager sets up the controller with the Manager.
+func (r *ConfigDeploymentReconciler) SetupWithManager(mgr ctrl.Manager) error {
+	/*
+		The `configMap` field must be indexed by the manager, so that we will be able to lookup `ConfigDeployments` by a referenced `ConfigMap` name.
+		This will allow for quickly answer the question:
+		- If ConfigMap _x_ is updated, which ConfigDeployments are affected?
+	*/
+
+	if err := mgr.GetFieldIndexer().IndexField(context.Background(), &appsv1.ConfigDeployment{}, configMapField, func(rawObj client.Object) []string {
+		// Extract the ConfigMap name from the ConfigDeployment Spec, if one is provided
+		configDeployment := rawObj.(*appsv1.ConfigDeployment)
+		if configDeployment.Spec.ConfigMap == "" {
+			return nil
+		}
+		return []string{configDeployment.Spec.ConfigMap}
+	}); err != nil {
+		return err
+	}
+
+	/*
+	As explained in the CronJob tutorial, the controller will first register the Type that it manages, as well as the types of subresources that it controls.
+	Since we also want to watch ConfigMaps that are not controlled or managed by the controller, we will need to use the `Watches()` functionality as well.
+
+	The `Watches()` function is a controller-runtime API that takes:
+	- A Kind (i.e. `ConfigMap`)
+	- A mapping function that converts a `ConfigMap` object to a list of reconcile requests for `ConfigDeployments`.
+	We have separated this out into a separate function.
+	- A list of options for watching the `ConfigMaps`
+	  - In our case, we only want the watch to be triggered when the ResourceVersion of the ConfigMap is changed.
+	 */
+
+	return ctrl.NewControllerManagedBy(mgr).
+		For(&appsv1.ConfigDeployment{}).
+		Owns(&kapps.Deployment{}).
+		Watches(
+			&source.Kind{Type: &corev1.ConfigMap{}},
+			handler.EnqueueRequestsFromMapFunc(r.findObjectsForConfigMap),
+			builder.WithPredicates(predicate.ResourceVersionChangedPredicate{}),
+		).
+		Complete(r)
+}
+
+/*
+	Because we have already created an index on the `configMap` reference field, this mapping function is quite straight forward.
+	We first need to list out all `ConfigDeployments` that use `ConfigMap` given in the mapping function.
+	This is done by merely submitting a List request using our indexed field as the field selector.
+
+	When the list of `ConfigDeployments` that reference the `ConfigMap` is found,
+	we just need to loop through the list and create a reconcile request for each one.
+	If an error occurs fetching the list, or no `ConfigDeployments` are found, then no reconcile requests will be returned.
+*/
+func (r *ConfigDeploymentReconciler) findObjectsForConfigMap(configMap client.Object) []reconcile.Request {
+	attachedConfigDeployments := &appsv1.ConfigDeploymentList{}
+	listOps := &client.ListOptions{
+		FieldSelector: fields.OneTermEqualSelector(configMapField, configMap.GetName()),
+		Namespace:     configMap.GetNamespace(),
+	}
+	err := r.List(context.TODO(), attachedConfigDeployments, listOps)
+	if err != nil {
+		return []reconcile.Request{}
+	}
+
+	requests := make([]reconcile.Request, len(attachedConfigDeployments.Items))
+	for i, item := range attachedConfigDeployments.Items {
+		requests[i] = reconcile.Request{
+			NamespacedName: types.NamespacedName{
+				Name:      item.GetName(),
+				Namespace: item.GetNamespace(),
+			},
+		}
+	}
+	return requests
+}