14. Using Dynamic Inventory in Ansible

In the previous article, we explored basic Inventory configuration methods, including how to define host information using a static hosts file. In this tutorial, we’ll dive deeper into Ansible’s dynamic Inventory functionality—which automatically generates host lists from cloud services, external databases, or other data sources. Dynamic Inventory offers exceptional flexibility, especially when managing infrastructure that changes frequently.

Overview of Dynamic Inventory

Dynamic Inventory is a host list generated by Ansible at runtime from external data sources. This means Ansible can fetch up-to-date host information each time a task runs—for example, by calling the AWS EC2 API, Google Compute Engine (GCE) API, or any other RESTful API.

Common use cases for dynamic Inventory include:

• Automatically discovering new instances in cloud environments
• Retrieving host information from a Configuration Management Database (CMDB)
• Integrating with container orchestration platforms such as Kubernetes

Configuring Dynamic Inventory

1. Creating a Dynamic Inventory Script

First, you need to write a script that outputs a JSON-formatted host list. Below is a simple example of a dynamic Inventory script returning two host groups:

#!/usr/bin/env python

import json

def get_inventory():
    inventory = {
        "web": {
            "hosts": ["web1.example.com", "web2.example.com"],
            "vars": {
                "http_port": 80,
            },
        },
        "db": {
            "hosts": ["db.example.com"],
            "vars": {
                "db_port": 5432,
            },
        },
    }
    return inventory

if __name__ == "__main__":
    print(json.dumps(get_inventory()))

Save the above code as dynamic_inventory.py and make it executable:

chmod +x dynamic_inventory.py

2. Configuring Ansible to Use the Dynamic Inventory Script

Next, instruct Ansible to use your dynamic Inventory script—by updating the ansible.cfg file:

[defaults]
inventory = /path/to/dynamic_inventory.py

Be sure to replace /path/to/dynamic_inventory.py with the actual path to your script.

3. Testing the Dynamic Inventory

Run the following command in your terminal to verify that the dynamic Inventory works correctly:

ansible-inventory --list

You should see the host list output in JSON format.

4. Using the Dynamic Inventory

Once configured, you can use the dynamic Inventory just like a static one. For example, run the following command to ping all hosts:

ansible all -m ping

Real-World Example: Dynamic Inventory with AWS EC2

Suppose you’re running a fleet of EC2 instances on AWS and want to manage them automatically using dynamic Inventory. You can leverage the boto3 library to fetch instance metadata. Here's an example script:

#!/usr/bin/env python

import json
import boto3

def get_inventory():
    ec2 = boto3.resource('ec2')
    instances = ec2.instances.filter(Filters=[{'Name': 'instance-state-name', 'Values': ['running']}])
    
    inventory = {
        'all': {
            'hosts': [],
        }
    }
    
    for instance in instances:
        inventory['all']['hosts'].append(instance.public_ip_address)
    
    return inventory

if __name__ == "__main__":
    print(json.dumps(get_inventory()))

This script uses the boto3 client to connect to AWS and retrieve the public IP addresses of all running EC2 instances. When you run ansible-inventory --list, those IPs will appear dynamically in the output.

Summary

In this article, we thoroughly examined how to use dynamic Inventory to flexibly manage Ansible host definitions. By writing simple Python scripts, you can significantly simplify infrastructure management—especially in dynamic cloud environments. Whether integrating with AWS, GCE, or other services, dynamic Inventory enables effortless discovery and management of cloud resources, greatly enhancing automation efficiency and operational convenience.

In the next article, we’ll explore grouping hosts and configuring host variables within Inventory—stay tuned!