[octavia-ingress-controller] Does not support a single entry point model, creates multiple LoadBalancers for multiple Ingress resources

[fakman0]

/kind feature

What happened:
When multiple Ingress resources are created in the Kubernetes cluster, each with the kubernetes.io/ingress.class: "openstack" annotation (or the corresponding ingressClassName), the octavia-ingress-controller creates a separate Octavia LoadBalancer for each Ingress resource. This results in multiple public IP addresses being allocated, with each IP corresponding to a single Ingress definition.

What you expected to happen:
I expected the octavia-ingress-controller to follow the "Single Entry Point" model, which is the standard behavior for most major ingress controllers (like NGINX).

The expected behavior is that the controller should manage a single Octavia LoadBalancer and a single public IP address. When new Ingress resources are created, the controller should read their rules and dynamically update the configuration of this single, existing LoadBalancer (by adding new listeners, pools, or rules) instead of creating a new one. This allows multiple applications and services to be exposed through one centralized ingress point, which is more cost-effective and easier to manage.

How to reproduce it:
Just apply this yaml file:

apiVersion: apps/v1
kind: Deployment
metadata:
  name: sample-app-deployment
spec:
  replicas: 1
  selector:
    matchLabels:
      app: sample-app
  template:
    metadata:
      labels:
        app: sample-app
    spec:
      containers:
      - name: nginx
        image: nginx:latest
        ports:
        - containerPort: 80
          name: http-port
        - containerPort: 8080
          name: grpc-port
---
apiVersion: v1
kind: Service
metadata:
  name: sample-app-service
spec:
  type: NodePort
  selector:
    app: sample-app
  ports:
  - name: http
    protocol: TCP
    port: 80
    targetPort: http-port
  - name: grpc
    protocol: TCP
    port: 8080
    targetPort: grpc-port
---
apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
  name: http-ingress
  annotations:
    octavia.ingress.kubernetes.io/backend-protocol: "HTTP"
    kubernetes.io/ingress.class: "openstack"
    octavia.ingress.kubernetes.io/internal: "false"
spec:
  rules:
  - host: "your.example.com"
    http:
      paths:
      - path: /
        pathType: Prefix
        backend:
          service:
            name: sample-app-service
            port:
              name: http
---
apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
  name: grpc-ingress
  annotations:
    octavia.ingress.kubernetes.io/backend-protocol: "GRPC"
    kubernetes.io/ingress.class: "openstack"
    octavia.ingress.kubernetes.io/internal: "false"
spec:
  rules:
  - host: "your.example.com"
    http:
      paths:
      - path: /grpc
        pathType: Prefix
        backend:
          service:
            name: sample-app-service
            port:
              name: grpc

My expectation here is that it is managed through a single loadbalancer, because the same domain name is just ingress rules that address different endpoints and protocols. However, for this scenario, 2 separate loadbalancers are created and I encounter 2 different floating IP addresses.

Anything else we need to know?:
The current behavior significantly increases the cost and complexity of using the controller. Exposing multiple services requires allocating multiple Floating IPs and managing multiple DNS records.

Environment:

• octavia-ingress-controller version: [v1.32.0]
• Kubernetes version: [v1.32.5+rke2r1]