Yükleniyor…

DevOps

End-to-End DevOps Pipeline with Terraform + Ansible + CI

In the modern DevOps world, just deploying an application is not enough. The entire infrastructure, configuration, and deployment processes are expected to be fully automated.

In this article, we'll walk through a real-world scenario: we'll provision infrastructure with Terraform, configure and deploy with Ansible, and run everything with a single git push using a GitHub Actions CI pipeline.

Scenario

The fully automated flow we're targeting works like this:

Developer pushes code to GitHub
CI pipeline is triggered
Terraform creates the server on AWS
Ansible connects to the server, installs Docker, and brings up the application
Application runs in production

Terraform defines the infrastructure; Ansible configures that infrastructure. The two tools complement each other perfectly: Terraform answers "what exists," while Ansible answers "how is it configured."

1. Infrastructure Provisioning with Terraform

We create an EC2 instance on AWS and expose its public IP. Ansible will use this IP to connect to the server.

hcl

1provider "aws" {

2 region = "eu-central-1"

5resource "aws_instance" "web" {

6 ami = "ami-0a5b2c0f7f9d"

7 instance_type = "t2.micro"

8 key_name = "devops-key"

10 tags = {

11 Name = "DevOps-Server"

12 }

13}

The output block passes the IP address to the pipeline for Ansible to use in the next steps.

hcl

1output "public_ip" {

2 value = aws_instance.web.public_ip

2. Configuration and Deployment with Ansible

After Terraform creates the server, Ansible takes over. It installs Docker and starts the application container.

yaml

1- hosts: web

2 become: yes

4 tasks:

5 - name: Install Docker

6 apt:

7 name: docker.io

8 state: present

9 update_cache: yes

11 - name: Start container

12 docker_container:

13 name: myapp

14 image: baykara/myapp:latest

15 ports:

16 - "80:5000"

17 state: started

The playbook is idempotent; running the same playbook multiple times does not change the result. This is critical for enabling safe re-deployments.

3. CI — GitHub Actions Pipeline

The GitHub Actions workflow that ties everything together. It is automatically triggered on every push to the main branch.

yaml

1name: Full DevOps Pipeline

3on:

4 push:

5 branches: [ "main" ]

7jobs:

8 deploy:

9 runs-on: ubuntu-latest

11 steps:

12 - name: Checkout repo

13 uses: actions/checkout@v4

15 - name: Install Terraform

16 uses: hashicorp/setup-terraform@v3

18 - name: Terraform Init

19 run: terraform init

20 working-directory: terraform

22 - name: Terraform Apply

23 run: terraform apply -auto-approve

24 working-directory: terraform

26 - name: Get IP address

27 run: echo "SERVER_IP=$(terraform output -raw public_ip)" >> $GITHUB_ENV

28 working-directory: terraform

30 - name: Install Ansible

31 run: sudo apt install -y ansible

33 - name: Create inventory

34 run: |

35 echo "[web]" > ansible/inventory

36 echo "$SERVER_IP ansible_user=ubuntu" >> ansible/inventory

38 - name: Deploy

39 run: ansible-playbook -i ansible/inventory ansible/deploy.yml

Conclusion

When these three tools come together, a true "push to deploy" experience emerges. The developer only writes code; provisioning the infrastructure, preparing the server, and deploying the application all happen automatically.

Terraform → Infrastructure (what exists?)
Ansible → Configuration (how is it set up?)
GitHub Actions → Orchestration (when and in what order?)

This pattern forms the foundation of a battle-tested DevOps infrastructure that scales from small startups to large enterprise teams.

All Articles

Yükleniyor…

Back to Blog

DevOps

End-to-End DevOps Pipeline with Terraform + Ansible + CI

In the modern DevOps world, just deploying an application is not enough. The entire infrastructure, configuration, and deployment processes are expected to be fully automated.

Scenario

The fully automated flow we're targeting works like this:

Developer pushes code to GitHub
CI pipeline is triggered
Terraform creates the server on AWS
Ansible connects to the server, installs Docker, and brings up the application
Application runs in production

1. Infrastructure Provisioning with Terraform

We create an EC2 instance on AWS and expose its public IP. Ansible will use this IP to connect to the server.

hcl

1provider "aws" {

2 region = "eu-central-1"

5resource "aws_instance" "web" {

6 ami = "ami-0a5b2c0f7f9d"

7 instance_type = "t2.micro"

8 key_name = "devops-key"

10 tags = {

11 Name = "DevOps-Server"

12 }

13}

The output block passes the IP address to the pipeline for Ansible to use in the next steps.

hcl

1output "public_ip" {

2 value = aws_instance.web.public_ip

2. Configuration and Deployment with Ansible

After Terraform creates the server, Ansible takes over. It installs Docker and starts the application container.

yaml

1- hosts: web

2 become: yes

4 tasks:

5 - name: Install Docker

6 apt:

7 name: docker.io

8 state: present

9 update_cache: yes

11 - name: Start container

12 docker_container:

13 name: myapp

14 image: baykara/myapp:latest

15 ports:

16 - "80:5000"

17 state: started

The playbook is idempotent; running the same playbook multiple times does not change the result. This is critical for enabling safe re-deployments.

3. CI — GitHub Actions Pipeline

The GitHub Actions workflow that ties everything together. It is automatically triggered on every push to the main branch.

yaml

1name: Full DevOps Pipeline

3on:

4 push:

5 branches: [ "main" ]

7jobs:

8 deploy:

9 runs-on: ubuntu-latest

11 steps:

12 - name: Checkout repo

13 uses: actions/checkout@v4

15 - name: Install Terraform

16 uses: hashicorp/setup-terraform@v3

18 - name: Terraform Init

19 run: terraform init

20 working-directory: terraform

22 - name: Terraform Apply

23 run: terraform apply -auto-approve

24 working-directory: terraform

26 - name: Get IP address

27 run: echo "SERVER_IP=$(terraform output -raw public_ip)" >> $GITHUB_ENV

28 working-directory: terraform

30 - name: Install Ansible

31 run: sudo apt install -y ansible

33 - name: Create inventory

34 run: |

35 echo "[web]" > ansible/inventory

36 echo "$SERVER_IP ansible_user=ubuntu" >> ansible/inventory

38 - name: Deploy

39 run: ansible-playbook -i ansible/inventory ansible/deploy.yml

Conclusion

Terraform → Infrastructure (what exists?)
Ansible → Configuration (how is it set up?)
GitHub Actions → Orchestration (when and in what order?)

This pattern forms the foundation of a battle-tested DevOps infrastructure that scales from small startups to large enterprise teams.

All Articles