Mattermost is working but the cluster is reconciling

[RomainMou]

Summary

When deploying Mattermost with the operator, Mattermost works fine but the cluster is reconciling and the operator logs that it failed to check ingress:

time="2023-09-15T09:40:43Z" level=info msg="[opr.controllers.Mattermost] Mattermost instance not healthy" Request.Name=mattermost Request.Namespace=staging fields.msg="failed to check ingress load balancer: expected one ingress, but found 0"
time="2023-09-15T09:40:49Z" level=info msg="[opr.controllers.Mattermost] Reconciling Mattermost" Request.Name=mattermost Request.Namespace=staging
time="2023-09-15T09:40:49Z" level=info msg="[opr.controllers.Mattermost] Mattermost instance not healthy" Request.Name=mattermost Request.Namespace=staging fields.msg="failed to check ingress load balancer: expected one ingress, but found 0"
time="2023-09-15T09:40:55Z" level=info msg="[opr.controllers.Mattermost] Reconciling Mattermost" Request.Name=mattermost Request.Namespace=staging
time="2023-09-15T09:40:55Z" level=info msg="[opr.controllers.Mattermost] Mattermost instance not healthy" Request.Name=mattermost Request.Namespace=staging fields.msg="failed to check ingress load balancer: expected one ingress, but found 0"

Steps to reproduce

Install Mattermost 9.0.0 (or 8.1.0) with Operator 1.20.1.
The issue occur when the operator and the cluster are in the same Namespace or different namespaces.

My ingress configuration:

  apiVersion: installation.mattermost.com/v1beta1
  kind: Mattermost
  metadata:
    name: "mattermost"
  spec:
    image: mattermost/mattermost-enterprise-edition
    imagePullPolicy: IfNotPresent
    version: 9.0.0
    size: 100users
    ingress:
      enabled: true
      host: "team.domain.tld"
      annotations:
        kubernetes.io/ingress.class: nginx
        cert-manager.io/cluster-issuer: letsencrypt-prod
        nginx.ingress.kubernetes.io/force-ssl-redirect: "true"
      ingressClass: nginx

Expected behavior

I expected the operator to find the ingress it created, and the cluster te be in an ok state.

Observed behavior (that appears unintentional)

The cluster is constantly reconciling even if it's working. The operator logs that it can't find the ingress.

Possible fixes

I guess this is an issue with this function:
https://github.com/mattermost/mattermost-operator/blob/3c7c810e5a3205b95640b3fdb779b6d1ffd2d813/pkg/mattermost/healthcheck/health_check.go#L166C1-L182C1