management of the config.river with the conversion of the config.yaml

[lbrule]

This fork will permit to realize a config.river of the flow mode thanks to the config.yaml of the static mode.

The objective is to deploy a grafana agent in mode flow with a dynamic configuration and not only with a config.river already configured before the deployement by ansible.

The solution that i have used it to convert the static format to the flow format is the following command:

• name: Create Grafana Agent River Config if flow mode for Grafana Agent
  ansible.builtin.shell: "AGENT_MODE=flow {{ grafana_agent_install_dir }}/grafana-agent convert -f static {{ grafana_agent_config_dir }}/{{ grafana_agent_config_filename }} -o {{ grafana_agent_config_dir }}/{{ grafana_agent_config_filename }}"
  notify: "restart grafana-agent"
  when: grafana_agent_provisioned_config_file | length == 0

I have made any tests and it is ok for me

Regards
Ludovic

[CLAassistant]

All committers have signed the CLA.

[ishanjainn]

Hey @lbrule , Thanks for adding this.

Also could you share a snippet of how you would use this in a playbook?

[lbrule]

/- hosts: all
tasks:
- name: Install Grafana Agent
ansible.builtin.include_role:
name: grafana.grafana.grafana_agent
vars:
grafana_agent_mode: flow
# Change config file on the host to .river
grafana_agent_config_filename: config.river
# Remove default flags
grafana_agent_flags_extra:
server.http.listen-addr: '0.0.0.0:12345'

[ishanjainn]

Sorry for the delay, LGTM
Will get it merged in next release

[ishanjainn]

@gardar any thoughts here?

[gardar]

How would this work on the second run of the ansible role when it has already converted the config?

[AdminAosCompagny]

Hello @gardar

The config.river is re generated always whene we run again the role. So every time we generate the config.yaml for the static mode and we do a conversion in order to obtain the config.river for the flow mode.

[ishanjainn]

@gardar is this okay from your aspect?