dce-addon/insight-agent chart has repeated operator.insight.io/managed-by field in template charts/kube-prometheus-stack/templates/exporters/core-dns/servicemonitor.yaml

[DonaldKellett]

dce-addon/insight-agent chart has repeated operator.insight.io/managed-by field in template charts/kube-prometheus-stack/templates/exporters/core-dns/servicemonitor.yaml which causes HelmRelease reconciliation failures with Flux v2 as shown in the screenshot below:

Chart details

• Repository: dce-addon (https://release.daocloud.io/chartrepo/addon)
• Chart name: insight-agent
• Chart version: 0.21.1 (affects 0.19.x as well)

Steps to reproduce

Unfortunately, it appears some templates rely on data from the running cluster so helm template doesn't work.

Install DCE5 Community with dce5-installer according to the official instructions, then run:

helm -n insight-system get manifest insight-agent > insight-agent.yaml.txt

Now inspect the output of the command in insight-agent.yaml.txt (attached below).

insight-agent.yaml.txt

In the YAML block with source charts/kube-prometheus-stack/templates/exporters/core-dns/servicemonitor.yaml, observe a duplicate operator.insight.io/managed-by field:

---
# Source: insight-agent/charts/kube-prometheus-stack/templates/exporters/core-dns/servicemonitor.yaml
apiVersion: monitoring.coreos.com/v1
kind: ServiceMonitor
metadata:
  name: insight-agent-kube-prometh-coredns
  namespace: insight-system
  labels:
    app: kube-prometheus-stack-coredns
    operator.insight.io/managed-by: insight
    
    app.kubernetes.io/managed-by: Helm
    app.kubernetes.io/instance: insight-agent
    app.kubernetes.io/version: "45.28.1"
    app.kubernetes.io/part-of: kube-prometheus-stack
    chart: kube-prometheus-stack-45.28.1
    release: "insight-agent"
    heritage: "Helm"
    operator.insight.io/managed-by: insight
spec:
  jobLabel: jobLabel
  
  selector:
    matchLabels:
      app: kube-prometheus-stack-coredns
      release: "insight-agent"
  namespaceSelector:
    matchNames:
      - "kube-system"
  endpoints:
  - port: http-metrics
    bearerTokenFile: /var/run/secrets/kubernetes.io/serviceaccount/token

For reference, here is the values.yaml file used (password REDACTED for security reasons):

global:
  exporters:
    auditLog:
      host: insight-opentelemetry-collector.insight-system.svc.cluster.local
    logging:
      host: 192.168.86.201
      password: REDACTED
      port: 31824
      scheme: https
      user: elastic
    metric:
      host: vminsert-insight-victoria-metrics-k8s-stack.insight-system.svc.cluster.local
kube-prometheus-stack:
  prometheus:
    prometheusSpec:
      replicas: 1
opentelemetry-collector:
  replicaCount: 1

[DonaldKellett]

On further inspection, it seems this issue occurs in multiple files servicemonitor.yaml under charts/kube-prometheus-stack/templates/exporters/ likely due to a common cause.