Question: How to enable rolling update of coreDNS pods when autoscaler is enabled

[kky-fury]

Hello all,

I am trying to update coreDNS image versions across different Kubernetes clusters.

I am using the helm chart version 1.31.0 with the following relevant values:

replicaCount: 6
rollingUpdate:
    maxUnavailable: 1
    maxSurge: 25%
...
autoscaler:
    enabled: true
    min: 4
    max: 100
podDisruptionBudget:
    minAvailable: 1

When I do the update all coreDNS pods are started at once without respecting the rollingUpdate constraints. I think it might be because of how replicas are defined in templates/deployment.yaml:

spec:
  {{- if and (not .Values.autoscaler.enabled) (not .Values.hpa.enabled) }}
  replicas: {{ .Values.replicaCount }}

Can you please suggest possible methods to do the update while respecting the rolling update constraints in this scenario.

Thank you for the help.