Issues with syncing custom PostgreSQL Helm chart settings in ArgoCD

[angelosleebos]

Issues with syncing custom PostgreSQL Helm chart settings in ArgoCD

Hi there,

I’ve created a custom Helm chart to manage PostgreSQL resources for the Zalando Postgres Operator, but I’m running into issues where some settings don’t seem to sync properly in ArgoCD. It’s unclear if the issue is related to the customizations, ArgoCD, or the operator itself.

Here are the key components of my setup. Basically, only the postgresql.yaml is relevant to my question, but for completeness, I've included the other components as well in case there are any comments on those.

---

PostgreSQL resource (templates/postgresql.yaml):

apiVersion: acid.zalan.do/v1
kind: postgresql
metadata:
  name: {{ .Release.Name }}-db
spec:
  {{- if .Values.connectionPooler }}
  connectionPooler:
    maxDBConnections: {{ .Values.connectionPooler.maxDBConnections }}
    mode: {{ .Values.connectionPooler.mode | default "transaction" }}
  {{- end }}
  databases:
  {{- range .Values.databases }}
    {{ . }}: admin
  {{- end }}
  {{- if .Values.nodeAffinity.lifecycle }}
  nodeAffinity: 
    requiredDuringSchedulingIgnoredDuringExecution:
      nodeSelectorTerms:
        - matchExpressions:
            - key: lifecycle
              operator: In
              values:
                - {{ .Values.nodeAffinity.lifecycle }}
  {{- end }}
  numberOfInstances: {{ .Values.numberOfInstances | default 2 }}
  postgresql:
    version: {{ .Values.version | quote }}
    {{- if .Values.parameters }}
    parameters: {{- toYaml .Values.parameters | nindent 6 }}
    {{- end }}
  resources: {{- toYaml .Values.resources | nindent 4 }}
  teamId: test
  users:
    admin:
      - superuser
      - createdb
    {{- if .Values.additionalUsers }}
    {{- range .Values.additionalUsers }}
    {{ . }}: []
    {{- end }}
    {{- end }}
  volume: {{- toYaml .Values.volume | nindent 4 }}

  {{- if .Values.walBackup.enabled }}
  enableLogicalBackup: true
  podEnvironmentConfigMap: "{{ .Release.Namespace }}/{{ .Release.Name }}-wal-config"
  serviceAccountName: {{ .Values.walBackup.serviceAccountName }}
  {{- end }}

---

ConfigMap for WAL-G (templates/wal-backup-configmap.yaml):

{{- if .Values.walBackup.enabled }}
apiVersion: v1
kind: ConfigMap
metadata:
  name: {{ .Release.Name }}-wal-config
  namespace: {{ .Release.Namespace }}
data:
  WALG_S3_PREFIX: "s3://{{ .Values.walBackup.s3Bucket }}"
  AWS_REGION: "{{ .Values.walBackup.awsRegion }}"
  USE_WALG_BACKUP: "true"
  USE_WALG_RESTORE: "true"
  BACKUP_SCHEDULE: "{{ .Values.walBackup.schedule }}"
  BACKUP_NUM_TO_RETAIN: "{{ .Values.walBackup.retention }}"
{{- end }}

---

ServiceAccount for WAL-G (templates/wal-backup-serviceaccount.yaml):

{{- if .Values.walBackup.enabled }}
apiVersion: v1
kind: ServiceAccount
metadata:
  name: {{ .Values.walBackup.serviceAccountName }}
  namespace: {{ .Release.Namespace }}
  annotations:
    eks.amazonaws.com/role-arn: "{{ .Values.walBackup.iamRoleArn }}"
{{- end }}

---

Part of values.yaml:

walBackup:
  enabled: true
  serviceAccountName: "wal-backup-sa"
  iamRoleArn: "arn:aws:iam::<account-id>:role/<iam-role-name>"
  s3Bucket: "<your-s3-bucket>"
  awsRegion: "eu-west-1"
  schedule: "30 00 * * *"
  retention: 7

---

Observed Behavior

• Some settings (e.g., podEnvironmentConfigMap, serviceAccountName) don’t seem to apply as expected.

---

Questions:

1. Are podEnvironmentConfigMap and serviceAccountName valid fields in the Postgresql CR spec? Are there known limitations for these fields?
2. Are there any operator-specific behaviors or configurations that could prevent these settings from applying correctly in ArgoCD?

[angelosleebos]

I figured out the backup configuration is supposed to be part of OperatorConfiguration.