[bitnami/rabbitmq-cluster-operator] metrics-endpoint serviceMonitor references missing port

[tolleiv]

Name and Version

bitnami/rabbitmq-cluster-operator 4.2.8

What architecture are you using?

amd64

What steps will reproduce the bug?

Release the chart with it's default values and monitoring enabled.

helm template rabbitmq-cluster-operator "oci://registry-1.docker.io/bitnamicharts/rabbitmq-cluster-operator" --version 4.2.8 --set clusterOperator.metrics.serviceMonitor.enabled=true --set clusterOperator.metrics.service.enabled=true --namespace=demo

The rendered resources include a ServiceMonitor with two endpoints:

---
# Source: rabbitmq-cluster-operator/templates/cluster-operator/servicemonitor.yaml
apiVersion: monitoring.coreos.com/v1
kind: ServiceMonitor
metadata:
  name: rabbitmq-cluster-operator
    ...
spec:
  ...
  endpoints:
    - port: http
    - port: metrics
---

where the related Deployment only has a single port:

---
# Source: rabbitmq-cluster-operator/templates/cluster-operator/deployment.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
  name: rabbitmq-cluster-operator
  ...
spec:
    ...
    spec:
      ...
      containers:
        - name: rabbitmq-cluster-operator
          ...
          ports:
            - name: http
              containerPort: 9782
              protocol: TCP
---

Are you using any custom parameters or values?

clusterOperator.metrics.serviceMonitor.enabled=true
clusterOperator.metrics.service.enabled=true

What is the expected behavior?

The ServiceMonitor should only reference existing ports.

What do you see instead?

The metrics Port is part of the ServiceMonitor without being present on the Deployment.

Additional information

The second port was introduce in dceeb50 - maybe a feature flag should be added or some more documentation.

[tolleiv]

Maybe also the author @basit9958 of the change dceeb50 could add their two cents on this one.

[github-actions]

This Issue has been automatically marked as "stale" because it has not had recent activity (for 15 days). It will be closed if no further activity occurs. Thanks for the feedback.

[tolleiv]

Before this get's stale, I'll try to come up with a pull request in the next days. Would be great to get input on this one beforehand.

[javsalgar]

Hi,

I believe that the metrics endpoint references the RabbitMQ instances created by the operator. It's not clear to me whether the selector actually matches the created instances.

[tolleiv]

Hi,

I believe that the metrics endpoint references the RabbitMQ instances created by the operator. It's not clear to me whether the selector actually matches the created instances.

The RabbitMQ instances do not have metrics endpoints either, those are named prometheus -> https://github.com/rabbitmq/cluster-operator/blob/HEAD/internal/resource/service.go#L240 - not sure what the initial idea behind the change was.

[carrodher]

Thank you for opening this issue and submitting the associated Pull Request. Our team will review and provide feedback. Once the PR is merged, the issue will automatically close.

Your contribution is greatly appreciated!