Ansible Event Driven Automation
Traditional playbooks start when a person or a schedule triggers them. Event Driven Ansible (EDA) reverses the idea: the automation listens for signals and reacts on its own. A smoke detector offers a good comparison. A person does not press a button to start the alarm, because the detector senses smoke and responds immediately. EDA watches monitoring tools, webhooks, and message queues, and launches playbooks the moment a matching event arrives.
Core Concepts
| Term | Meaning |
|---|---|
| Source | Plugin that receives events, such as a webhook or a Kafka topic |
| Event | A piece of data describing something that happened |
| Condition | A test that decides whether an event matters |
| Action | The response, such as running a playbook or a job template |
| Rule | A pairing of conditions with actions |
| Rulebook | A YAML file that holds sources and rules |
Diagram: From Event to Fix
┌───────────────┐ alert ┌─────────────────┐
│ Monitoring │──────────▶│ Source │
│ tool │ │ (webhook plugin)│
└───────────────┘ └────────┬────────┘
│ event
▼
┌─────────────────┐
│ Rulebook │
│ condition match?│
└───┬─────────┬───┘
yes│ │no
▼ ▼
┌──────────────┐ ignore
│ run_playbook │
│ restart_app │
└──────┬───────┘
▼
┌──────────────┐
│ Server fixed │
└──────────────┘
Install Ansible Rulebook
python3 -m pip install ansible-rulebook ansible-runner
ansible-rulebook --version
The tool needs Java 17 or later, because the rules engine runs on the Drools library.
Write a Rulebook
This rulebook listens on a web address and restarts a service when an alert reports high memory.
---
- name: React to monitoring alerts
hosts: all
sources:
- ansible.eda.webhook:
host: 0.0.0.0
port: 5000
rules:
- name: Restart app on high memory
condition: event.payload.alert == "high_memory" and event.payload.service == "shop"
action:
run_playbook:
name: restart_shop.yml
- name: Log every other alert
condition: event.payload.alert is defined
action:
debug:
msg: "Alert received: {{ event.payload.alert }}"
The Response Playbook
# restart_shop.yml
- hosts: webservers
become: true
tasks:
- name: Restart the shop service
ansible.builtin.service:
name: shop
state: restarted
Start the Rulebook and Send a Test Event
ansible-rulebook --rulebook alerts.yml -i inventory.ini --verbose
curl -X POST http://localhost:5000/endpoint \
-H "Content-Type: application/json" \
-d '{"alert": "high_memory", "service": "shop"}'
The terminal running the rulebook prints a message that the rule matched, and the playbook starts.
Popular Event Sources
- ansible.eda.webhook accepts HTTP posts from any tool
- ansible.eda.alertmanager receives Prometheus alerts
- ansible.eda.kafka reads messages from a Kafka topic
- ansible.eda.url_check polls a web address and reports status changes
- ansible.eda.file_watch reports file changes
Available Actions
| Action | Result |
|---|---|
| run_playbook | Runs a local playbook |
| run_job_template | Starts a template in AWX or Automation Controller |
| run_module | Runs one module |
| set_fact | Stores data for later rules |
| post_event | Creates a new event for other rules |
| debug | Prints information |
Design Tips
- Begin with low risk responses such as ticket creation and notifications
- Add throttling so repeated alerts do not trigger endless restarts
- Keep every response playbook idempotent
- Log each action for later review
- Require human approval for destructive actions
Try This
Build the sample rulebook with only the debug action. Start it and send three test events with different alert names through curl. Watch which rules fire. Add a throttle section to the main rule and send the same event twice within a minute to observe the filtering.
Summary
Event Driven Ansible listens for events and responds through rulebooks. Sources receive data from monitoring tools and message systems, conditions filter events, and actions run playbooks or modules. Start with safe responses, add throttling, and keep playbooks idempotent to build self-healing infrastructure that reacts in seconds.
