20. Best Practices

20.1. Use Source Control

While Tower supports playbooks stored directly on the Tower server, best practice is to store your playbooks, roles, and any associated details in source control. This way you have an audit trail describing when and why you changed the rules that are automating your infrastructure. Plus, it allows for easy sharing of playbooks with other parts of your infrastructure or team.

20.2. Ansible file and directory structure

Please review the Ansible best practices from the Ansible documentation at If creating a common set of roles to use across projects, these should be accessed via source control submodules, or a common location such as /opt. Projects should not expect to import roles or content from other projects.


Playbooks should not use the vars_prompt feature, as Tower does not interactively allow for vars_prompt questions. If you must use vars_prompt, refer to and make use of the Surveys functionality of Tower.

Jobs run in Tower use the playbook directory as the current working directory, although jobs should be coded to use the playbook_dir variable rather than relying on this.

20.3. Use Dynamic Inventory Sources

If you have an external source of truth for your infrastructure, whether it is a cloud provider or a local CMDB, it is best to define an inventory sync process and use Tower’s support for dynamic inventory (including cloud inventory sources and custom inventory scripts). This ensures your inventory is always up to date.


With the release of Ansible Tower 2.4.0, edits and additions to Inventory host variables now persist beyond an inventory sync as long as --overwrite_vars is not set. To have inventory syncs behave as they did before, it is now required that both --overwrite and --overwrite_vars are set.

20.4. Variable Management for Inventory

Keeping variable data along with the objects in Tower (see the inventory editor) is encouraged, rather than using group_vars/ and host_vars/. If you use dynamic inventory sources, Tower can sync such variables with the database as long as the Overwrite Variables option is not set.

20.5. Autoscaling

Using the “callback” feature to allow newly booting instances to request configuration is very useful for auto-scaling scenarios or provisioning integration.

20.6. Larger Host Counts

Consider setting “forks” on a job template to larger values to increase parallelism of execution runs. For more information on tuning Ansible, see the Ansible blog.

20.7. Continuous integration / Continuous Deployment

For a Continuous Integration system, such as Jenkins, to spawn an Tower job, it should make a curl request to a job template, or use the Tower CLI tool. The credentials to the job template should not require prompting for any particular passwords. Using the API to spawn jobs is covered in the Tower API guide.