TIL: Library Helm charts share templates, not workloads

Every chart I work on ends up with the same handful of helper templates: standard labels, a fullname helper, the same resource-limit boilerplate. I used to copy them from chart to chart. Library charts are Helm’s answer to that.

A library chart is a chart with type: library in its Chart.yaml:

apiVersion: v2
name: common
version: 0.1.0
type: library

Instead of manifests, it only ships named templates — the define blocks you’d normally keep in _helpers.tpl:

{{- define "common.labels" -}}
app.kubernetes.io/name: {{ .Chart.Name }}
app.kubernetes.io/managed-by: {{ .Release.Service }}
{{- end }}

An application chart pulls it in as a regular dependency:

# the app chart's Chart.yaml
dependencies:
  - name: common
    version: 0.1.0
    repository: file://../common

and then calls the helpers like its own:

metadata:
  labels:
    {{- include "common.labels" . | nindent 4 }}

The library chart itself deploys nothing. It has no rendered manifests of its own, and Helm refuses to install it standalone — it exists purely to be depended on.