ec2ansible is a cli tool written in Python (2.7+) for generating realtime AWS EC2 inventory for Ansible.

There is an official script doing the similar thing. But usually when we are provisioning a server, we only care about its role, eg. web server, database server, proxy server, etc, instead of its availability zone, AMI ID, nor instance type. ec2ansible groups your EC2 servers into Ansible inventory hierarchically based on the Role tag associated with them.

Asumming we have two main types of servers in the us-east-1 region: web servers and worker servers. And we have totally 5 servers as follow:

• Apache
• Nginx reverse proxy
• HAProxy reverse proxy
• Gearman worker
• Celery worker

Firstly we will add a Role tag for each of them with the following values:

• web_apache
• web_proxy_nginx
• web_proxy_haproxy
• worker_gearman
• worker_celery

Here is a sample grouping generated by ec2ansible:

use1 (us-east-1)
└─── web
     │    web_apache
     └─── web_proxy
          │    web_proxy_nginx
          │    web_proxy_haproxy
└─── worker
          │    worker_gearman
          │    worker_celery

ec2ansible is available via pypi and can be installed with easy_install or pip:

pip install ec2ansible

1. Save ec2ansible-sample.ini to ~/.ec2ansible and modify it if needed
2. Specify /usr/local/bin/ec2ansible as the inventory for Ansible

ansible web_apache -i /usr/local/bin/ec2ansible -u ec2-user -m ping

or set in ansible.cfg

hostfile = /usr/local/bin/ec2ansible