Closed chgl closed 2 years ago
To follow best-practices when deploying more than one replica of the server, we should allow configuring the pod affinity, nodeSelector and toleration settings. See https://kubernetes.io/docs/concepts/scheduling-eviction/assign-pod-node/
A fairly simple, yet flexible implementation from the chart-users point of view would be the following:
{{- with .Values.nodeSelector }} nodeSelector: {{- toYaml . | nindent 8 }} {{- end }} {{- with .Values.affinity }} affinity: {{- toYaml . | nindent 8 }} {{- end }} {{- with .Values.tolerations }} tolerations: {{- toYaml . | nindent 8 }} {{- end }}
Additionally, adding a PDB (https://kubernetes.io/docs/tasks/run-application/configure-pdb/) is helpful from an "ops" perspective as well. See e.g. https://github.com/bitnami/charts/blob/master/bitnami/airflow/templates/web/poddisruptionbudget.yaml
Also clsoed by #65 🚀
To follow best-practices when deploying more than one replica of the server, we should allow configuring the pod affinity, nodeSelector and toleration settings. See https://kubernetes.io/docs/concepts/scheduling-eviction/assign-pod-node/
A fairly simple, yet flexible implementation from the chart-users point of view would be the following:
Additionally, adding a PDB (https://kubernetes.io/docs/tasks/run-application/configure-pdb/) is helpful from an "ops" perspective as well. See e.g. https://github.com/bitnami/charts/blob/master/bitnami/airflow/templates/web/poddisruptionbudget.yaml