New to KubeDB? Please start here.

AppBinding

What is AppBinding

An AppBinding is a Kubernetes CustomResourceDefinition(CRD) which points to an application using either its URL (usually for a non-Kubernetes resident service instance) or a Kubernetes service object (if self-hosted in a Kubernetes cluster), some optional parameters and a credential secret. To learn more about AppBinding and the problems it solves, please read this blog post: The case for AppBinding.

If you deploy a database using KubeDB, AppBinding object will be created automatically for it. Otherwise, you have to create an AppBinding object manually pointing to your desired database.

KubeDB uses Stash to perform backup/recovery of databases. Stash needs to know how to connect with a target database and the credentials necessary to access it. This is done via an AppBinding.

AppBinding CRD Specification

Like any official Kubernetes resource, an AppBinding has TypeMeta, ObjectMeta and Spec sections. However, unlike other Kubernetes resources, it does not have a Status section.

An AppBinding object created by KubeDB for ZooKeeper database is shown below,

apiVersion: appcatalog.appscode.com/v1alpha1
kind: AppBinding
metadata:
  annotations:
    kubectl.kubernetes.io/last-applied-configuration: |
      {"apiVersion":"kubedb.com/v1alpha2","kind":"ZooKeeper","metadata":{"annotations":{},"name":"zk-cluster","namespace":"demo"},"spec":{"podTemplate":{"spec":{"containers":[{"name":"zookeeper","resources":{"requests":{"cpu":"720m","memory":"846Mi"}}}]}},"replicas":3,"serviceTemplates":[{"alias":"primary","spec":{"type":"LoadBalancer"}}],"storage":{"accessModes":["ReadWriteOnce"],"resources":{"requests":{"storage":"1Gi"}},"storageClassName":"standard"},"terminationPolicy":"WipeOut","version":"3.9.1"}}      
  creationTimestamp: "2024-05-02T10:02:45Z"
  generation: 2
  labels:
    app.kubernetes.io/component: database
    app.kubernetes.io/instance: zk-cluster
    app.kubernetes.io/managed-by: kubedb.com
    app.kubernetes.io/name: zookeepers.kubedb.com
  name: zk-cluster
  namespace: demo
  ownerReferences:
    - apiVersion: kubedb.com/v1alpha2
      blockOwnerDeletion: true
      controller: true
      kind: ZooKeeper
      name: zk-cluster
      uid: 20e00408-abf1-470b-a049-bdf272b3e994
  resourceVersion: "3548"
  uid: 8fd15549-ab9c-4523-b85d-77275f620bb5
spec:
  appRef:
    apiGroup: kubedb.com
    kind: ZooKeeper
    name: zk-cluster
    namespace: demo
  clientConfig:
    service:
      name: zk-cluster
      port: 2181
      scheme: http
  secret:
    name: zk-cluster-auth
  type: kubedb.com/zookeeper
  version: 3.9.1

Here, we are going to describe the sections of an AppBinding crd.

AppBinding Spec

An AppBinding object has the following fields in the spec section:

spec.type

spec.type is an optional field that indicates the type of the app that this AppBinding is pointing to. Stash uses this field to resolve the values of TARGET_APP_TYPE, TARGET_APP_GROUP and TARGET_APP_RESOURCE variables of BackupBlueprint object.

This field follows the following format: <app group>/<resource kind>. The above AppBinding is pointing to a zookeeper resource under kubedb.com group.

Here, the variables are parsed as follows:

spec.secret

spec.secret specifies the name of the secret which contains the credentials that are required to access the database. This secret must be in the same namespace as the AppBinding.

This secret must contain the following keys:

spec.appRef

appRef refers to the underlying application. It has 4 fields named apiGroup, kind, name & namespace.

spec.clientConfig

spec.clientConfig defines how to communicate with the target database. You can use either a URL or a Kubernetes service to connect with the database. You don’t have to specify both of them.

You can configure following fields in spec.clientConfig section:

• spec.clientConfig.url

  spec.clientConfig.url gives the location of the database, in standard URL form (i.e. [scheme://]host:port/[path]). This is particularly useful when the target database is running outside of the Kubernetes cluster. If your database is running inside the cluster, use spec.clientConfig.service section instead.

Note that, attempting to use a user or basic auth (e.g. user:password@host:port) is not allowed. Stash will insert them automatically from the respective secret. Fragments ("#…") and query parameters ("?…") are not allowed either.

• spec.clientConfig.service

  If you are running the database inside the Kubernetes cluster, you can use Kubernetes service to connect with the database. You have to specify the following fields in spec.clientConfig.service section if you manually create an AppBinding object.

  • name : name indicates the name of the service that connects with the target database.
  • scheme : scheme specifies the scheme (i.e. http, https) to use to connect with the database.
  • port : port specifies the port where the target database is running.

• spec.clientConfig.insecureSkipTLSVerify

  spec.clientConfig.insecureSkipTLSVerify is used to disable TLS certificate verification while connecting with the database. We strongly discourage to disable TLS verification during backup. You should provide the respective CA bundle through spec.clientConfig.caBundle field instead.

• spec.clientConfig.caBundle

  spec.clientConfig.caBundle is a PEM encoded CA bundle which will be used to validate the serving certificate of the database.

Next Steps

• Learn how to use KubeDB to manage various databases here.
• Want to hack on KubeDB? Check our contribution guidelines.