I want to do some testing across multiple hosts (of an OpenStack platform) and I’d really like to base this on Ansible to take advantage of the inventory management.
However, I really don’t want the standard ansible-playbook output, even in check mode, I’d like something where I can craft the output.
I’d like to avoid reinventing the wheel, just like I’d like to avoid making bad implementation choices but I’m surprised not to find an existing framework as it seems such a good use case for Ansible.
So not having found anything it strikes me that I could
- use the Ansible API … seems to be a no no as there seem to be no guarantees of stability across releases
- write a python wrapper which invokes ansible via the command-line and parses the output … not very nice
I’m sure I’ve missed something obvious.
Any pointers?
Thanks,
Mike.