Helm support for single or multiple values files

[telemaco]

SUMMARY

Could be interesting that the helm module will support single or multiple values files.
Helm command itself support it adding "-f values.yaml -f values2.yaml"

ISSUE TYPE

• Feature Idea

COMPONENT NAME

• helm

ADDITIONAL INFORMATION

[orjan]

Since the module is taking values we can combine the files with Ansible without having to use the built in support in helm for multiple files. I'll guess something like this might work?

---
- hosts: all
  vars:
    default_vars: "{{ lookup('template', 'default.yaml') | from_yaml }}"
    override_vars: "{{ lookup('template', 'production.yaml') | from_yaml }}"
  tasks:
    - name: Combine values
      debug:
        msg: "{{ default_vars | combine(override_vars) }}"

https://github.com/ansible-collections/community.kubernetes/blob/master/plugins/modules/helm.py#L150

[loomsen]

I have a similar use case, where I have a common values file, and I'd like to define the overrides in the helm task in the playbook itself, to make it more readable. For example, prometheus could be installed with alertmanager in stg and prd, but without in dev. It would be nice to have a common file, and have tasks in the playbooks like:

        - name: install prometheus
          helm:
            chart: stable/prometheus
            name: prometheus
            namespace: monitoring
            values_file: /path/to/common/values.yml
            values:
                 alertmanager:
                   enabled: false

Would be nice to have a way to combine those, like for example a values_file parameter. I have tried to combine it in different ways, but none of them worked. So now either I have to have different template files for each environment, even though they might just differ in one line, or I would have massively long and unreadable playbooks, all due to the huge prometheus values.yaml when you define custom scrape configs and whatnot.

So what I ended up doing in the end, which imo is not very nice:

        - name: install prometheus
          helm:
            chart_ref: stable/prometheus
            name: prometheus
            release_namespace: monitoring
            values: "{{ default_values | combine(overrides) }}"
          vars:
            default_values: "{{ lookup('template', '{{ playbook_dir }}/helm_values/prometheus.yml') | from_yaml }}"
            overrides:
              alertmanager:
                ingress:
                  hosts:
                    - alertmanager.{{ cluster.dns.internal_domain }}
                  tls:
                    - hosts:
                      - alertmanager.{{ cluster.dns.internal_domain }}

[telemaco]

Sorry for the close and reopen. I just want to comment that I've similar situation than @loomsen.

[ggee]

Same situation. Above works but not clean.