CI/CD Pipeline for NGINX on Docker — Phase 4: Ansible & Dynamic Inventory

Phase 4 configures the Terraform-provisioned EC2 instance and deploys the Docker container — without ever hardcoding a host IP.

What I Built

1. AWS dynamic inventory — Instead of a static hosts file, Ansible queries the AWS API at runtime to discover which EC2 instances exist, their IPs, and their tags. If Terraform creates a new instance, Ansible finds it automatically on the next run.
2. Configuration playbook — An Ansible playbook that SSHs into the discovered instances and:
   • Installs Docker
   • Pulls the NGINX image from Docker Hub
   • Runs the container with the correct port bindings (80, 443)
   • Configures the container to restart on reboot

Why Dynamic Inventory

Static inventory breaks as soon as infrastructure changes — a new instance gets a different IP and the playbook targets the wrong host. Dynamic inventory solves this by querying AWS each time. The pipeline doesn't care which instance exists, only that it can find and configure whatever Terraform created.

What's Next

Phase 5 wires everything together with Jenkins — a single Git push triggers Terraform, then Ansible, then a smoke test confirming the deployment is live.