Connecting Applications with Operator-backed Services
Introduction
The goal of the Service Binding Operator is to enable application authors to import an application
and run it on OpenShift with operator-backed services such as databases, without having to perform
manual configuration of secrets, configmaps, etc.
In order for the Service Binding Operator to bind an application to a backing service, the backing
service operator must specify the information required by the application to bind to the
operator's service. The information must be specified in the operator's OLM (Operator Lifecycle
Manager) descriptor from which it will be extracted to bind the application to the operator. The
information could be specified in the "status" and/or "spec" section of the OLM in plaintext or as
a reference to a secret.
In order to make an imported application (for example, a NodeJS application) connect to a backing
services (for example, a database):
Here is an example of the bind-able operator OLM Descriptor -- in this case for a PostgreSQL
database backing operator:
---
[...]
statusDescriptors:
description: Name of the Secret to hold the DB user and password
displayName: DB Password Credentials
path: dbCredentials
x-descriptors:
- urn:alm:descriptor:io.kubernetes:Secret
- urn:alm:descriptor:servicebindingrequest:env:object:secret:user
- urn:alm:descriptor:servicebindingrequest:env:object:secret:password
description: Database connection IP address
displayName: DB IP address
path: dbConnectionIP
x-descriptors:
- urn:alm:descriptor:servicebindingrequest:env:attribute
Quick Start
Clone the repository and run make local in an existing kube:admin openshift CLI session.
Alternatively, install the operator using:
cat <<EOS |kubectl apply -f -
---
apiVersion: operators.coreos.com/v1
kind: OperatorSource
metadata:
name: redhat-developer-operators
namespace: openshift-marketplace
spec:
type: appregistry
endpoint: https://quay.io/cnr
registryNamespace: redhat-developer
EOS
Example Scenario
In this example there are 2 roles:
- Cluster Admin - Installs the operators to the cluster
- Application Developer - Imports a Node.js application, creates a DB instance, creates a request to bind the application and DB (to connect the DB and the application).
Cluster Admin
The cluster admin needs to install 2 operators into the cluster:
- Service Binding Operator
- Backing Service Operator
A Backing Service Operator that is "bind-able," in other
words a Backing Service Operator that exposes binding information in secrets, config maps, status, and/or spec
attributes. The Backing Service Operator may represent a database or other services required by
applications. We'll use postgresql-operator to
demonstrate a sample use case.
Install the Service Binding Operator using an OperatorSource
Apply the following OperatorSource:
cat <<EOS |kubectl apply -f -
---
apiVersion: operators.coreos.com/v1
kind: OperatorSource
metadata:
name: redhat-developer-operators
namespace: openshift-marketplace
spec:
type: appregistry
endpoint: https://quay.io/cnr
registryNamespace: redhat-developer
EOS
Then navigate to the Operators->OperatorHub in the OpenShift console and in the Other category select the Service Binding Operator operator

and install a stable version.
This makes the ServiceBindingRequest custom resource available, that the application developer will use later.
Install the DB operator using an OperatorSource
Apply the following OperatorSource:
cat <<EOS |kubectl apply -f -
---
apiVersion: operators.coreos.com/v1
kind: OperatorSource
metadata:
name: db-operators
namespace: openshift-marketplace
spec:
type: appregistry
endpoint: https://quay.io/cnr
registryNamespace: pmacik
EOS
Then navigate to the Operators->OperatorHub in the OpenShift console and in the Database category select the PostgreSQL Database operator

and install a stable version.
This makes the Database custom resource available, that the application developer will use later.
Application Developer
Create a namespace called service-binding-demo
The application and the DB needs a namespace to live in so let's create one for them:
cat <<EOS |kubectl apply -f -
---
kind: Namespace
apiVersion: v1
metadata:
name: service-binding-demo
EOS
Import an application
In this example we will import an arbitrary Node.js application.
In the OpenShift Console switch to the Developer perspective. (Make sure you have selected the service-binding-demo project). Navigate to the +ADD page from the menu and then click on the [Import from Git] button. Fill in the form with the following:
Git Repo URL = https://github.com/pmacik/nodejs-rest-http-crud
Project = service-binding-demo
Application->Create New Application = NodeJSApp
Name = nodejs-app
Builder Image = Node.js
Create a route to the application = checked
and click on the [Create] button.
Notice, that during the import no DB config was mentioned or requestd.
When the application is running navigate to its route to verify that it is up. Notice that in the header it says (DB: N/A). That means that the application is not connected to a DB and so it should not work properly. Try the application's UI to add a fruit - it causes an error proving that the DB is not connected.
Create a DB instance for the application
Now we utilize the DB operator that the cluster admin has installed. To create a DB instance just create a Database custom resource in the service-binding-demo namespace called db-demo:
cat <<EOS |kubectl apply -f -
---
apiVersion: postgresql.baiju.dev/v1alpha1
kind: Database
metadata:
name: db-demo
namespace: service-binding-demo
spec:
image: docker.io/postgres
imageName: postgres
dbName: db-demo
EOS
Set labels on the application
Now the we need to set arbitrary labels on the application's DeploymentConfig in order for the Service Binding Operator to be able to find the application.
The labels are:
connects-to=postgres - indicates that the application needs to connect to a PostgreSQL DB
environment=demo - indicates the demo environment - it narrows the search
kubectl patch dc nodejs-app -p '{"metadata": {"labels": {"connects-to": "postgres", "environment":"demo"}}}'
Express an intent to bind the DB and the application
Now the only thing that remains is to connect the DB and the application. We let the Service Binding Operator to 'magically' do the connection for us.
Create the following ServiceBindingRequest:
cat <<EOS |kubectl apply -f -
---
apiVersion: apps.openshift.io/v1alpha1
kind: ServiceBindingRequest
metadata:
name: binding-request
namespace: service-binding-demo
spec:
applicationSelector:
matchLabels:
connects-to: postgres
environment: demo
resourceKind: DeploymentConfig
backingServiceSelector:
resourceKind: postgresql.baiju.dev
resourceVersion: v1alpha1
resourceRef: db-demo
mountPathPrefix: “”
EOS
There are 2 parts in the request:
applicationSelector - used to search for the application based on the labels that we set earlier and the resourceKind of the application to be a DeploymentConfig.
backingServiceSelector - used to find the backing service - our operator-backed DB instance called db-demo.
That causes the application to be re-deployed.
Once the new version is up, go the application's route to check the UI. In the header you can see (DB: db-demo) which indicates that the application is connected to a DB and its name is db-demo. Now you can try the UI again but now it works!
When the ServiceBindingRequest was created the Service Binding Operator's controller injected the DB connection information as specified in the OLM descriptor below, into the
application's DeploymentConfig as environment variables via an intermediate Secret called binding-request:
spec:
template:
spec:
containers:
- envFrom:
- secretRef:
name: binding-request
ServiceBindingRequestStatus
ServiceBindingRequestStatus depicts the status of the Service Binding operator. More info: https://github.com/kubernetes/community/blob/master/contributors/devel/api-conventions.md#spec-and-status
| Field |
Description |
| BindingStatus |
The binding status of Service Binding Request |
| Secret |
The name of the intermediate secret |
That's it, folks!