Ansible playbooks can generate quite a lot of console data. Add the -v (I do it by default !) and you have a quite a lot to scroll through. Often times, one tends to feel the need for a better reporting mechanism — one which is easy to scour for specific task status, whilst having the luxury to extract additional debug info for the same if needed. Also, it would be great if this aid is available for playbook runs across time, i.e., for older playbook runs as well. This would be especially beneficial when running ansible-based CI suites, like @OpenEBS, where the application deployments, test setup & execution is driven by Ansible.
Sounds like a perfect requirement for a UI-based solution, doesn’t it ?
A lot of people using Ansible for medium-large scale deployments are known to use Tower, Rundeck or the opensource alternative semaphore to achieve this (Tower has a lifelong self-support trial license without a few features for under 10 nodes). There is also Foreman, which is great in environments where a hybrid tool set — puppet with ansible, or chef with ansible etc.., are used. Most of these are workflow management tools that do more than just what we desired above, i.e,. playbook recording. These tools provide a centralised management capability wherein inventory, users, task/play scheduling, notifications can all be controlled from a dashboard. Now, it is also possible to integrate these with other popular CI tools like Jenkins (watch this).
However, if there is already a system in place to perform some of the things these tools do (dynamic inventories, playbook triggers/scheduling) while not really needing others (access control, graphs), and you are only looking for the ability to store and analyse playbook runs via UI, then ARA (Ansible Run Analysis) is your tool.
Heavily used by openstack community in their CI projects, ARA is built to just “record” playbook runs (Read ARA’s manifesto, to understand more about its narrow focus). It does this via an ansible callback plugin to store run details into a database and a web interface to visualise the database.
While it uses sqlite and an embedded webserver, respectively, for these purposes, you could even customise it to use mysql & Apache. The UI includes nifty features like host fact lookup, playbook params page, search filter, property based sort and link to code snippets !
A nice video explaining the web interface, with playbooks from openstack-ansible project is here
In addition to these, ARA also provides
- Couple of ansible modules for persisting & viewing some user data that one may want to view on the browser-based UI (like an ansible “fact”, but for visualisation purposes :P).
- A CLI to query the database (While I haven’t found much use for it till now, it is useful to custom create some reports)
ARA follows the same support cycle as the upstream Ansible community, and is under active development (See github)
Here is an ansible role we have written that you could use to quickly setup ARA on your ubuntu box 🙂
We @OpenEBS, have found great benefit in using this tool and would happily recommend it for the usecases discussed. !!
