Description of problem:

When multiple CLFs define an HTTP input with the same name, the HTTP input service is reconciled endlessly.

How reproducible: Always

Steps to Reproduce:

1. Create multiple CLFs with the same receiver input names.

• CLF 1

  apiVersion: logging.openshift.io/v1
  kind: ClusterLogForwarder
  metadata:
    name: http-only1
    namespace: openshift-logging
  spec:
    serviceAccountName: test-sa
    inputs:
    - name: http-audit
      receiver:
        http:
          format: kubeAPIAudit
          port: 8080
    outputs:
      - name: my-splunk
        type: splunk
        url: https://splunk-service.splunk.svc.cluster.local:8088
        secret:
          name: splunk-secret
        tls:
          insecureSkipVerify: true
    pipelines:
     - name: app-logs
       inputRefs:
        - http-audit
       outputRefs:
        - my-splunk
  

• CLF2

  apiVersion: logging.openshift.io/v1
  kind: ClusterLogForwarder
  metadata:
    name: http-only2
    namespace: openshift-logging
  spec:
    serviceAccountName: test-sa
    inputs:
    - name: http-audit
      receiver:
        http:
          format: kubeAPIAudit
          port: 8080
    outputs:
      - name: my-splunk
        type: splunk
        url: https://splunk-service.splunk.svc.cluster.local:8088
        secret:
          name: splunk-secret
        tls:
          insecureSkipVerify: true
    pipelines:
     - name: app-logs
       inputRefs:
        - http-audit
       outputRefs:
        - my-splunk
  

Apply both and check `Networking -> Services`

Actual results:

The receiver input service is continuously reconciled.

Expected results:

There should be a service for each CLF

Additional Info:

Affects Logging 5.8 and Logging 5.9