diff --git a/docs/kubernetes_development.md b/docs/kubernetes_development.md
index 3d0484e..b88e1cb 100644
--- a/docs/kubernetes_development.md
+++ b/docs/kubernetes_development.md
@@ -73,24 +73,23 @@ kubectl apply -f roles.yaml
 apiVersion: apps/v1
 kind: Deployment
 metadata:
-  name: calls-offloader
+  name: calls-offloader-deployment
   labels:
-    app: calls
+    app.kubernetes.io/name: calls-offloader
 spec:
-  replicas: 1
+  replicas: 2
   selector:
     matchLabels:
-      app: calls
+      app.kubernetes.io/name: calls-offloader
   template:
     metadata:
       labels:
-        app: calls
+        app.kubernetes.io/name: calls-offloader
     spec:
       serviceAccountName: calls-offloader-service-account
-      hostNetwork: true # This is needed purely for testing so that it's easier to connect from the local MM instance.
       containers:
       - name: calls-offloader
-        image: calls-offloader:master # Testing a local image, should point to the official registry when running in prod.
+        image: calls-offloader:dev-d94cee0 # Testing a local image, should point to the official registry when running in prod.
         ports:
         - containerPort: 4545
         env:
@@ -98,44 +97,74 @@ spec:
             valueFrom:
               fieldRef:
                 fieldPath: metadata.namespace
-          - name: DEV_MODE # This is needed for testing.
+          - name: DEV_MODE # This is only needed for testing. Should be removed for production use.
             value: "true"
           - name: LOGGER_ENABLEFILE
             value: "false"
           - name: JOBS_APITYPE
             value: "kubernetes"
-          - name: API_SECURITY_ALLOWSELFREGISTRATION
+          - name: JOBS_MAXCONCURRENTJOBS
+            value: "1"
+          - name: API_SECURITY_ALLOWSELFREGISTRATION # This should only be set to true if running the service inside a private network.
             value: "true"
           - name: LOGGER_CONSOLELEVEL
             value: "DEBUG"
 ```
 
-Finally we create the deployment using the YAML file above:
+We then create the deployment using the YAML file above:
 
 ```sh
 kubectl apply -f deployment.yaml
 ```
 
-### Verify pod is running correctly
+### Create service
 
-To verify the deployment is running correctly:
+```yaml
+apiVersion: v1
+kind: Service
+metadata:
+  name: calls-offloader-service
+spec:
+  type: NodePort
+  selector:
+    app.kubernetes.io/name: calls-offloader
+  ports:
+    - protocol: TCP
+      port: 4545
+      targetPort: 4545
+      nodePort: 30045
+```
+
+Finally we create a `NodePort` service to expose the pods to the node:
 
 ```sh
-kubectl get deployment calls-offloader
+kubectl apply -f service.yaml
 ```
 
+### Verify pods are running correctly
+
+First verify the deployment was created correctly:
+
 ```sh
-kubectl logs -l app=calls
+kubectl get deployment -l app.kubernetes.io/name=calls-offloader
 ```
 
-### Check pod IP address
+Then verify the pods are running by checking the logs:
 
-Finally we find the IP address for the pod to be used from the Calls side to connect to the service (*Job Service URL*):
+```sh
+kubectl logs -l app.kubernetes.io/name=calls-offloader
+```
+
+### Expose service URL
+
+Finally we need to expose the service so that it can be accessed from the host.
 
 ```sh
-kubectl get pods -l app=calls -o wide
+minikube service calls-offloader-service --url
 ```
 
+The returned URL should be used as the value of the *Job service URL* in the Calls configuration settings.
+
 > **_Note_**
 >
 > The host's IP (e.g. 192.168.49.1) needs to be configured as *ICE Host Override* on the Calls side to get connectivity to calls from within pod to work.