This content originally appeared on DEV Community and was authored by A-LPHARM
Kubernetes has always been evolving, and this progress provides more flexibility and extensibility to developers and operations teams. one feature I came across is Custom Resource Definition (CRDs). this allows users to create their own new resource type within Kubernetes extending its API, this flexibility is key in an ecosystem as diverse as that of cloud-native applications. Developers can create a custom resource with all the rules just like your built-in Kubernetes resources such as Pods, Rbac, Replicasets, API Services, Then your Kubernetes API server leverages all the custom resources.
** What are Custom Resource Definition (CRD)?**
A resource
is an endpoint in the Kubernetes API that stores a collection of API objects of a certain Kind and custom resource
is an extension of the Kubernetes API that is not necessarily available in a default Kubernetes installation.
Custom Resource Definition
is a kubernetes resource that allows you to define custom resource
in the Kubernetes API.
What happens when you Create a Custom Resource Definition?
when you create the CRD definition kubernetes validates it against the schema defined in CRD. this includes checking the apiVersion
, kind
, metadata
, spec and all rules specified in the openAPIV3Schema
once its validated it is accepted and stored in the cluster etcd, the etcd storage ensures that the state of your custom resources is persistent and can be retrieved and managed across the cluster.
then the API server
creates a new RESTful API endpoint for your custom resource. example if you create a CRD for a custom resource called pdfdocs
in the alpharm.henry.com
API group, the API server will create endpoints suh as:
-
/apis/alpharm.com/v1/pdfdocs
-
/apis/alpharm.com/v1/namespaces/{namespace}/henrycustomresources/{name}
these endpoints can be used to perform CRUD operations.
The resources are then monitored by controllers that are responsible for taking the information in the resource, once you have a custom controller or operator it manages the custom resource and then take appropriate to ensure that the desired state of your custom resources is maintained.
features to consider when CRD is created
- Version:
v1
it identifies the version of the custom resource. - Kind:
pdfdoc
this is typically the name of the custom resource. -
spec.names
defines how to describe your custom resource. -
metadata.name
this is the name of the CRD - Group:
alpharm.henry.com
it is usually attached to the crd resource name - plural:
pdfdocs
- Scope: namespaced this determines if the CR can be created in a namespace or globally.
-
Spec.names.kind
: this determines the kind of CR to be created using CRD like deployment, cronjob.
Creating the Custom Resource Definition (CRD)
create a file with crd_manifest.yaml
apiVersion: apiextensions.k8s.io/v1
kind: CustomResourceDefinition
metadata:
name: pdfdocs.alpharm.henry.com #plural_name.the_group_name
spec:
group: alpharm.henry.com
names:
kind: pdfdoc
plural: pdfdocs
shortNames:
- pd
scope: Namespaced # this defines where the custom resources instances will be created, namespaced or globally
versions:
- name: v1
served: true #this determines if this version v1 is used and will be served, if false no one uses it
storage: true #this dertermines if this version will be stored in etcd
schema:
openAPIV3Schema: #lets create our resource here
type: object
description: 'APIVersion defines the versioned schema of this representation
of an object. Servers should convert recognized schemas to the latest
internal value, and may reject unrecognized values. More info: https://git.k8s.io/community/contributors/devel/api-conventions.md#resources'
properties:
apiVersion: #our typical kubernetes resource format apiversion, kind, metadata, spec
type: string
kind:
type: string
description: 'Kind is a string value representing the REST resource this
object represents. Servers may infer this from the endpoint the client
submits requests to. Cannot be updated. In CamelCase. More info: https://git.k8s.io/community/contributors/devel/api-conventions.md#types-kinds'
metadata:
type: object
spec:
type: object
properties:
documentName:
type: string #uses array, boolean, integer, number, object, string
text:
type: string
then create the Custom Resource Definitions
$ kubectl apply -f crd.yaml
customresourcedefinition.apiextensions.k8s.io/pdfdocs.alpharm.henry.com created
$ kubectl get crd
NAME CREATED AT
pdfdocs.alpharm.henry.com 2024-07-30T12:37:28Z
when we run this to get the documents
$ kubectl get pdfdocs
No resources found in default namespace.
this signify that the pdf
is stored but there aren't any documents available in the pdfdocs
lets create a resource with our cluster resource definition
apiVersion: alpharm.henry.com/v1
kind: pdfdoc
metadata:
name: my-document
spec:
documentName: build
text: |
we are creating a document to test our first crd
so this can be fun to try out
then we create the resource
$ kubectl apply -f test-manifest.yaml
pdfdoc.alpharm.henry.com/my-document created
so we can view it
Custom resources consume storage space in the same way that ConfigMaps
do. Creating too many custom resources may overload your API server's storage space.
Now lets create Your Custom Controllers
Lets set up a Go environment for the controller create a directory and initalize a Go module:
mkdir k8s-controller
cd k8s-controller
go mod init k8s-controller
then you add the dependencies
go get k8s.io/apimachinery@v0.22.0 k8s.io/client-go@v0.22.0 sigs.k8s.io/controller-runtime@v0.9.0
next, we will write a file main.go and add this command.
main.go
package main
import (
"context"
"flag"
"log"
"os"
"os/signal"
"path/filepath"
"syscall"
metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
"k8s.io/apimachinery/pkg/apis/meta/v1/unstructured"
"k8s.io/apimachinery/pkg/runtime"
"k8s.io/apimachinery/pkg/runtime/schema"
"k8s.io/apimachinery/pkg/watch"
"k8s.io/client-go/dynamic"
"k8s.io/client-go/rest"
"k8s.io/client-go/tools/cache"
"k8s.io/client-go/tools/clientcmd"
"k8s.io/client-go/util/homedir"
)
func main() {
var kubeconfig string
if home := homedir.HomeDir(); home != "" {
kubeconfig = filepath.Join(home, ".kube", "config")
}
// Allow the kubeconfig file to be specified via a flag
flag.StringVar(&kubeconfig, "kubeconfig", kubeconfig, "absolute path to the kubeconfig file")
flag.Parse()
config, err := clientcmd.BuildConfigFromFlags("", kubeconfig)
if err != nil {
log.Println("Falling back to in-cluster config")
config, err = rest.InClusterConfig()
if err != nil {
log.Fatalf("Failed to get in-cluster config: %v", err)
}
}
dynClient, err := dynamic.NewForConfig(config)
if err != nil {
log.Fatalf("Failed to create dynamic client: %v", err)
}
pdfdoc := schema.GroupVersionResource{Group: "alpharm.henry.com", Version: "v1", Resource: "pdfdocs"}
informer := cache.NewSharedIndexInformer(
&cache.ListWatch{
ListFunc: func(options metav1.ListOptions) (runtime.Object, error) {
return dynClient.Resource(pdfdoc).Namespace("").List(context.Background(), options)
},
WatchFunc: func(options metav1.ListOptions) (watch.Interface, error) {
return dynClient.Resource(pdfdoc).Namespace("").Watch(context.Background(), options)
},
},
&unstructured.Unstructured{},
0,
cache.Indexers{},
)
informer.AddEventHandler(cache.ResourceEventHandlerFuncs{
AddFunc: func(obj interface{}) {
log.Println("Add event detected:", obj)
},
UpdateFunc: func(oldObj, newObj interface{}) {
log.Println("Update event detected:", newObj)
},
DeleteFunc: func(obj interface{}) {
log.Println("Delete event detected:", obj)
},
})
stop := make(chan struct{})
defer close(stop)
go informer.Run(stop)
if !cache.WaitForCacheSync(stop, informer.HasSynced) {
log.Fatalf("Timeout waiting for cache sync")
}
log.Println("Custom Resource Controller started successfully")
sigCh := make(chan os.Signal, 1)
signal.Notify(sigCh, syscall.SIGINT, syscall.SIGTERM)
<-sigCh
}
the file that was set up defines the path to your kubeconfig file and i set it at the default /.kube/config
but users can also specify the kubeconfig file path. this Go program sets up a basic kubernetes custom resource controller that listens for changes to my custom resource type pdfdocs. it uses the client-go library to interact with kubernetes, with the ability to use your local kubeconfig file, allowing for updates, add, delete of events for our custom resource.
Dockerize the controller
we create a docker image from this Go program, create a file dockerfile
# Use an official Golang image to build the Go application
FROM golang:1.22.5 AS build
# Set the working directory inside the container
WORKDIR /app
# Copy the go.mod and go.sum files and download dependencies
COPY go.mod go.sum ./
RUN go mod download
# Copy the rest of the application source code
COPY . .
# Build the Go application
RUN CGO_ENABLED=0 GOOS=linux go build -a -installsuffix cgo -o k8s-controller .
# Use a minimal base image for the final container
FROM alpine:3.14
# Copy the built Go binary from the builder stage
COPY --from=build /app/k8s-controller /usr/local/bin/k8s-controller
# Set the entrypoint to the Go application
ENTRYPOINT ["/usr/local/bin/k8s-controller"]
then build the image and push to your repository
docker build -t henriksin1/k8s-controller:v1 .
#then run
docker push henriksin1/k8s-controller:v1
the docker repo: https://hub.docker.com/repository/docker/henriksin1/k8s-controller
Set-up your Role-Based Access control
firstly, create your service account that your controller will use
apiVersion: v1
kind: ServiceAccount
metadata:
name: k8s-controller-sa
namespace: default
deploy the service account
kubectl apply -f 1-serviceaccout.yaml
next, create your cluster role and clusterrolebinding for the necessary permissions and, a clusterrolebinding to bind the clusterrole to the service account
apiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRole
metadata:
name: k8s-controller-role
rules:
- apiGroups: ["alpharm.henry.com"]
resources: ["pdfdocs"]
verbs: ["get", "list", "watch", "create", "update", "patch", "delete"]
deploy the clusterrole account
kubectl apply -f 2-clusterrole.yaml
ClusterRoleBinding:
apiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRoleBinding
metadata:
name: k8s-controller-binding
subjects:
- kind: ServiceAccount
name: k8s-controller-sa
namespace: default
roleRef:
kind: ClusterRole
name: k8s-controller-role
apiGroup: rbac.authorization.k8s.io
deploy the clusterrolebinding resource
kubectl apply -f 3-clusterrolebinding.yaml
Create your deployment and specify the service account name for the image to use.
apiVersion: apps/v1
kind: Deployment
metadata:
name: k8s-controller
labels:
app: k8s-controller
spec:
replicas: 1
selector:
matchLabels:
app: k8s-controller
template:
metadata:
labels:
app: k8s-controller
spec:
serviceAccountName: k8s-controller-sa
containers:
- name: k8s-controller
image: henriksin1/k8s-controller:v1
args: []
deploy the image
kubectl apply -f 4-deployment.yaml
once all resources are running confirm
you can log the pods and check
kubectl logs pod/k8s-controller-675b96c777-cld42
2024/07/31 12:30:42 Falling back to in-cluster config
2024/07/31 12:30:42 Add event detected: &{map[apiVersion:alpharm.henry.com/v1 kind:pdfdoc metadata:map[annotations:map[kubectl.kubernetes.io/last-applied-configuration:{"apiVersion":"alpharm.henry.com/v1","kind":"pdfdoc","metadata":{"annotations":{},"name":"my-document","namespace":"default"},"spec":{"documentName":"build","text":"we are creating a document to test our first crd\nso this can be fun to try out\n"}}
] creationTimestamp:2024-07-31T09:44:20Z generation:1 managedFields:[map[apiVersion:alpharm.henry.com/v1 fieldsType:FieldsV1 fieldsV1:map[f:metadata:map[f:annotations:map[.:map[] f:kubectl.kubernetes.io/last-applied-configuration:map[]]] f:spec:map[.:map[] f:documentName:map[] f:text:map[]]] manager:kubectl-client-side-apply operation:Update time:2024-07-31T09:44:20Z]] name:my-document namespace:default resourceVersion:1315 uid:3fbbf363-aae8-4916-9be5-95b8311c3bdb] spec:map[documentName:build text:we are creating a document to test our first crd
so this can be fun to try out
]]}
2024/07/31 12:30:42 Custom Resource Controller started successfully
now lets modify our pdfdocs
$ kubectl edit pd my-document
pdfdoc.alpharm.henry.com/my-document edited
[ec2-user@ip-172-31-20-185 k8s-controller]$ kubectl logs pod/k8s-controller-675b96c777-cld42
2024/07/31 12:30:42 Falling back to in-cluster config
2024/07/31 12:30:42 Add event detected: &{map[apiVersion:alpharm.henry.com/v1 kind:pdfdoc metadata:map[annotations:map[kubectl.kubernetes.io/last-applied-configuration:{"apiVersion":"alpharm.henry.com/v1","kind":"pdfdoc","metadata":{"annotations":{},"name":"my-document","namespace":"default"},"spec":{"documentName":"build","text":"we are creating a document to test our first crd\nso this can be fun to try out\n"}}
] creationTimestamp:2024-07-31T09:44:20Z generation:1 managedFields:[map[apiVersion:alpharm.henry.com/v1 fieldsType:FieldsV1 fieldsV1:map[f:metadata:map[f:annotations:map[.:map[] f:kubectl.kubernetes.io/last-applied-configuration:map[]]] f:spec:map[.:map[] f:documentName:map[] f:text:map[]]] manager:kubectl-client-side-apply operation:Update time:2024-07-31T09:44:20Z]] name:my-document namespace:default resourceVersion:1315 uid:3fbbf363-aae8-4916-9be5-95b8311c3bdb] spec:map[documentName:build text:we are creating a document to test our first crd
so this can be fun to try out
]]}
2024/07/31 12:30:42 Custom Resource Controller started successfully
2024/07/31 15:19:06 Update event detected: &{map[apiVersion:alpharm.henry.com/v1 kind:pdfdoc metadata:map[annotations:map[kubectl.kubernetes.io/last-applied-configuration:{"apiVersion":"alpharm.henry.com/v1","kind":"pdfdoc","metadata":{"annotations":{},"name":"my-document","namespace":"default"},"spec":{"documentName":"build","text":"we are creating a document to test our first crd\nso this can be fun to try out\n"}}
] creationTimestamp:2024-07-31T09:44:20Z generation:2 managedFields:[map[apiVersion:alpharm.henry.com/v1 fieldsType:FieldsV1 fieldsV1:map[f:metadata:map[f:annotations:map[.:map[] f:kubectl.kubernetes.io/last-applied-configuration:map[]]] f:spec:map[.:map[] f:documentName:map[]]] manager:kubectl-client-side-apply operation:Update time:2024-07-31T09:44:20Z] map[apiVersion:alpharm.henry.com/v1 fieldsType:FieldsV1 fieldsV1:map[f:spec:map[f:text:map[]]] manager:kubectl-edit operation:Update time:2024-07-31T15:19:06Z]] name:my-document namespace:default resourceVersion:29595 uid:3fbbf363-aae8-4916-9be5-95b8311c3bdb] spec:map[documentName:build text:this is a article about customresourcedefinition what do your think
so this can be fun to try out
Conclusion
Custom Resource Definitions (CRDs) empower Kubernetes users to extend the platform's capabilities, enabling the management of custom application resources using Kubernetes' declarative API. By leveraging CRDs, you can create custom controllers and operators to automate complex workflows and integrate seamlessly with external tools and services. Whether you're managing databases, custom application configurations, or automated workflows, CRDs provide the flexibility and power to tailor Kubernetes to your specific needs.
Feel free to experiment with CRDs and explore how they can simplify and enhance the management of your applications within Kubernetes.
see you'all next time
If you found this blog insightful and dive deeper into topics like AWS cloud, Kubernetes, and cloud native projects or anything related, check out my linkedin page: https://www.linkedin.com/in/emeka-henry-uzowulu-38900088/
and also, my github: https://github.com/A-LPHARM/
feel free to share your thoughts and ask any questions
for references:
Custom Resources | Kubernetes
Controllers | Kubernetes
This content originally appeared on DEV Community and was authored by A-LPHARM
A-LPHARM | Sciencx (2024-07-31T19:48:23+00:00) Understanding the Extension of Kubernetes APIs with Custom Resource Definition. Retrieved from https://www.scien.cx/2024/07/31/understanding-the-extension-of-kubernetes-apis-with-custom-resource-definition/
Please log in to upload a file.
There are no updates yet.
Click the Upload button above to add an update.