Allow passing custom arguments to the autoscaler

This can be useful if you have CoreDNS deployments that only serve a
subset of nodes, and you want to use --nodelabels on the autoscaler to
have it scale according to just those nodes.

Signed-off-by: Sarah Hodne <sarah@circleci.com>

Author: sarahhodne

charts/coredns/Chart.yaml

@@ -1,6 +1,6 @@
 apiVersion: v2
 name: coredns
-version: 1.16.7
+version: 1.17.0
 appVersion: 1.8.7
 home: https://coredns.io
 icon: https://coredns.io/images/CoreDNS_Colour_Horizontal.png

charts/coredns/README.md

@@ -113,6 +113,7 @@ The command removes all the Kubernetes components associated with the chart and
 | `autoscaler.max`                                | Max size of replicaCount                                                                                                                  | 0  (aka no max)                                             |
 | `autoscaler.includeUnschedulableNodes`          | Should the replicas scale based on the total number or only schedulable nodes                                                             | `false`                                                     |
 | `autoscaler.preventSinglePointFailure`          | If true does not allow single points of failure to form                                                                                   | `true`                                                      |
+| `autoscaler.customFlags`                        | A list of custom flags to pass into cluster-proportional-autoscaler                                                                       | (no args)                                                   |
 | `autoscaler.image.repository`                   | The image repository to pull autoscaler from                                                                                              | k8s.gcr.io/cluster-proportional-autoscaler-amd64            |
 | `autoscaler.image.tag`                          | The image tag to pull autoscaler from                                                                                                     | `1.8.1`                                                     |
 | `autoscaler.image.pullPolicy`                   | Image pull policy for the autoscaler                                                                                                      | IfNotPresent                                                |

charts/coredns/templates/deployment-autoscaler.yaml

@@ -86,4 +86,7 @@ spec:
           - --target=Deployment/{{ default (include "coredns.fullname" .) .Values.deployment.name }}
           - --logtostderr=true
           - --v=2
+        {{- if .Values.autoscaler.customFlags }}
+{{ toYaml .Values.autoscaler.customFlags | indent 10 }}
+        {{- end }}
 {{- end }}

charts/coredns/values.yaml

@@ -258,6 +258,11 @@ autoscaler:
   # If true does not allow single points of failure to form
   preventSinglePointFailure: true
 
+  ## Optionally specify some extra flags to pass to cluster-proprtional-autoscaler.
+  ## Useful for e.g. the nodelabels flag.
+  # customFlags:
+  #   - --nodelabels=topology.kubernetes.io/zone=us-east-1a
+
   image:
     repository: k8s.gcr.io/cluster-proportional-autoscaler-amd64
     tag: "1.8.1"