Since a number of our customers have had questions for us about the recent announcements around AWX [https://www.ansible.com/blog/upcoming-changes-to-the-awx-project/], we wanted to share our views on the current/future state of the AWX project as of Q4 2024, and make sure that everyone is making a fully-informed near-term decision that’s best for their organizations and goals with infrastructure automation.
Because Red Hat Ansible Automation Platform is rapidly moving toward a services-based architecture (long overdue for those of us who have been working under the hood)– and because AWX (as a project and as code) directly depends on AAP– you should be aware that AWX is no longer going to continue to be released as it has been historically.
Going forward, any AWX users who wish to upgrade will need to somehow bring together almost two-dozen upstream projects to try to even approach what is handled by a basic AAP deployment running out of a single installation (never mind enterprise components like Event Driven Ansible controller and Private Automation Hub and Automation Analytics– which are increasingly critical yet NOT available to replicate via upstream derivatives, with no plans to do so in the future).
20+ Discrete Community Upstream Projects that Red Hat brings together in a unified solution for AAP customers:
ansible-builder
ansible-test
ansible-language-server
Ansible Core, pylibSSH
Ansible Runner
AWX components now split out: awx-operator credential plugins, auth plugins, UI
ansible-navigator
ansible-lint
vscode-ansible
Content Collections, ansible-sign
Ansible Receptor
ansible-galaxy, Galaxy NG, Galaxy NG Operator, Hub UI
So if any organization is going to choose to be on AWX in the future, they will have to answer questions for themselves such as:
1) Where are we actually going to be getting any AWX software updates and bug fixes from?
2) How are we going to tell our leadership and business that our AWX deployment is secure? Especially when there is no process for certifying its source packages, etc?
3) Who’s going to actually do the updating of AWX now? Level Up is generally cautious on sharing any “FTE estimates needed to support open software project integrations”, but we’ve seen credible calculations that look like upwards of four FTE’s with different skill sets (heavy in Python development and DBA and security) to be across the 20+ discrete projects shown in above columns, assuming a 500 node deployment. So if a reasonable head count estimate is four, whatever the number of Python developers, etc. is for you, it’s not going to be zero– therefore, the idea that any customer will not need to make any type of significant staffing investment to support a stable AWX deployment? Probably optimistic.
In conclusion, like most classic open source projects, AWX has always been “free” like a puppy. However, AWX is now going to be “free” like 20+ puppies. Meanwhile, there’s Red Hat Ansible Automation Platform, which is intended to trade a small fraction of your overall IT budget for a massive amount of time savings, as well as unique security and reliability features which are simply not available in the open.