Using ARA Records Ansible Within Projects

2 minute read

While exploring some options to provide a level of reporting capabilities as part of a project I figured using ARA would be an awesome idea. My goal was to provide a way to include ARA as part of Git Repositories(projects) in which the consumers of those projects had the ability to gain visibility into reporting of Ansible executions. The idea would be to have ARA use a folder within the project such as logging/.ara which would become part of the project which could be committed for historical reasons. Which then could be served up from anywhere once the components of ARA are committed.

Sounds cool right? Well at least to me it seems to make a lot of sense. But as I started looking at the options of using ARA and configuring my ansible.cfg file which is within each project already I soon realized for some reason this would not work as I expected. So if we look at what the ansible.cfg file might look like as what I would imagine would work:




NOTE: The above information under defaults was obtained by running the following command: python -m ara.setup.ansible.

So what I found from the above is that ultimately ARA does not like for some reason defining dir=./.ara but I could use the full working directory and it was happy. Well that is not going to work as I will never know exactly where a consumer of these projects might actually be running Ansible from. In addition to this, the settings under [defaults] would most likely be different in each and every scenario as well. What if I put this under CI/CD? Who knows what that would look like and it surely could be different each and every time as well.

So how did I finally solve this? In all reality it is rather quite simple. And it will definitely work out much better IMHO as maybe the consumer may not want to always generate some ARA reporting output from their Ansible executions. So, to solve this, I would just simply add the following shell script as part of each project and let the consumer decide whether or not they wanted to generate ARA reporting. With that being said, the following is all it would take to make this achievable.

#!/usr/bin/env bash

ANSIBLE_CALLBACK_PLUGINS=$(python -m ara.setup.callback_plugins)
ANSIBLE_ACTION_PLUGINS=$(python -m ara.setup.action_plugins)
ANSIBLE_LIBRARY=$(python -m ara.setup.library)

export ARA_DIR

Now when a consumer would like to generate ARA reporting they can simply just execute the following source before executing their Ansible playbook. ARA would then use the current working directory and logging/.ara and then dynamically define ANSIBLE_CALLBACK_PLUGINS, ANSIBLE_ACTION_PLUGINS, and ANSIBLE_LIBRARY no matter which environment ran from assuming that ARA has been installed.

This would also play nicely when running through CI/CD processes as well so this should solve the requirement that I was aiming for!

Hope you enjoy and maybe someone else might find some use of this.


Leave a comment