[Bug]: Duplicate port name "healthcheck" in entity-operator

[eli-gc]

Bug Description

Seeing these duplicate port name warnings in entity-operator. This is because the topic-operator and user-operator containers within the entity-operator pod both define ports with name "healthcheck"

containers:
    - name: topic-operator
      image: quay.io/strimzi/operator:0.47.0
      args:
        - /opt/strimzi/bin/topic_operator_run.sh
      ports:
        - name: healthcheck
          containerPort: 8080
          protocol: TCP
...
      - name: user-operator
        image: quay.io/strimzi/operator:0.47.0
        args:
          - /opt/strimzi/bin/user_operator_run.sh
        ports:
          - name: healthcheck
            containerPort: 8081
            protocol: TCP

Steps to reproduce

Deploy entity operator with both user-operator and topic-operator.
I saw the warning when I did kubectl rollout restart

Expected behavior

It doesn't really seem to be affecting current functionality but it is probably best practice to name the ports distinct names. I think it would be affected if services or network policies reference the "healthcheck" port name.

Strimzi version

0.47.0

Kubernetes version

1.35.4

Installation method

Helm chart

Infrastructure

AKS

Configuration files and logs

No response

Additional context

No response

[scholzj]

Where exactly did you see the warnings and can you share them?

I think using unique port name makes sense. But it impacts the metrics scraping etc. So I think this needs to be carefully considered as it might break things / require users to update their metrics configuration..

[eli-gc]

@scholzj I see the warnings when I do kubectl rollout restart on the kafka entity-operator, so for example:

kubectl rollout restart deploy/kafka-entity-operator

[eli-gc]

This is the warning:
Warning: spec.template.spec.containers[1].ports[0]: duplicate port name "healthcheck" with spec.template.spec.containers[0].ports[0], services and probes that select ports by name will use spec.template.spec.containers[0].ports[0]

[scholzj]

Ok, thanks.

I'm not sure I understand why are you doing a rollout on the entity operator managed by the Cluster Operator. But I guess that does not matter that much.

[eli-gc]

We have istio sidecars in the pod so we needed to rollout restart for some upgrades there.

[eli-gc]

It still restarted though. Doesn't prevent me from restarting, but it might affect other things that need to select port by name. I don't think it really affects currently functionality but could be a gotcha in the future.

[katheris]

Discussed on community call on 28.5.26: Sounds like although this isn't affecting functionality it should be a straight forward fix so we should fix it. @eli-gc would you be interested in picking this up?

[eli-gc]

@katheris Sure I can do that.