You already have the network ready (VPC, public subnet, Internet Gateway, and routes). Now let's add the main character: an EC2 server inside that subnet, accessible from the internet. Remember everything from Chapter 4 about instances; now we'll write it in Terraform.

On top of the network from the previous subchapter, we add an EC2 instance:

Remember that every instance starts from an AMI (subchapter 4.2). Instead of manually writing an AMI ID (which changes over time and by region), it's best to automatically find the most recent one with a data block:

What is a data block? It's an important new concept: while resource creates something, a data block (data source) queries existing information in AWS without creating anything. Here we ask AWS "what is the most recent Amazon Linux 2023 AMI?" and save the answer to use it. It's very useful to avoid writing hardcoded values that expire.

To make the code flexible (subchapter 10.1), we define some variables:

Using t3.micro (subchapter 4.1) is ideal for learning, because it's included in the AWS free tier.

Now the main resource. We place it in our public subnet:

Let's analyze the connections:

• ami → reference to the data source from step 1 (the AMI found).
• instance_type → the variable from step 2.
• subnet_id → reference to the public subnet from subchapter 12.1. This places the server in our network, in the public zone.
• vpc_security_group_ids → reference to the Security Group (the firewall), which we will create in the next subchapter.

Each reference creates a dependency: Terraform will create the subnet and Security Group first, and then the instance.

Often you want the server to do something as soon as it starts, like installing a web server. For that, there is user_data: a script that runs automatically the first time the instance starts.

This script: updates the system, installs the Apache web server (httpd), starts it, and creates a welcome page. When the instance is ready, it will serve a website with that message.

What we just achieved: with user_data, the server self-configures at birth. You don't need to connect to install anything manually. This is key for automation and for autoscaling (Chapter 13): each new server prepares itself. (For complex configurations, it's better to create a ready-made AMI, remember subchapter 4.2, but user_data is perfect to start.)

Right now we have:

If you ran plan now, Terraform would complain that the Security Group (aws_security_group.web) we referenced is missing and hasn't been created yet. We'll solve that in the next subchapter.

• A data block (data source) queries existing information in AWS without creating anything (we use it to find the most recent AMI automatically, avoiding hardcoded IDs that expire).
• The instance is placed in the subnet with subnet_id, referencing the public subnet from the previous chapter.
• Using variables (like tipo_instancia) makes the code flexible; t3.micro is ideal for the free tier.
• user_data is a script that the server runs at startup, perfect for self-configuration (installing software, etc.) without manual intervention.
• References between resources create dependencies and the automatic creation order.

In the next subchapter, we will create the missing Security Group (firewall) and assign the server a fixed Elastic IP.