Is there a reason why ansible.cfg, which is an ini file, isn’t merged in precedence by Ansible when the file is found in different locations, but instead Ansible uses the first one it finds (in some order)? It’s very useful to have ansible.cfg in the root of your Ansible project because you can set things like roles_path and library, but it’s bad if you need to configure something differently which depends on the machine where Ansible will be executed. A good example is the forks, which you want to bump up on management hosts which have a lot of CPU and RAM, and decrease on smaller ones.
There isn’t a workaround for this currently, because Ansible doesn’t do merging. Is there a reason for it?
What could also be a good solution is to have an .include /etc/ansible.cfg statement in ansible.cfg, something like systemd has for its unit files, which would allow you to do something like this in the project ansible.cfg:
`
.include /etc/ansible/ansible.cfg
[defaults]
library = ./library
roles_path = ./roles
`
Which would basically mean use whatever it is in /etc/ansible/ansible.cfg and override library and roles_path.
What do you think about this?