Infrastructure as code with Terraform: a practical guide

Sat, 10 Sep 2022 00:00:00 +0000

Why infrastructure as code matters

Managing infrastructure manually through web consoles or ad-hoc scripts creates problems that pile up over time: inconsistent environments, undocumented changes, impossible rollbacks, and the classic “it works on my machine” extended to entire servers.

Infrastructure as Code (IaC) fixes this by treating infrastructure like application code: it’s written, versioned, reviewed, tested, and applied through automated workflows. The benefits show up right away:

Reproducibility: Spin up identical environments in minutes, not days.
Version control: Every infrastructure change goes through a PR with code review.
Documentation by default: The code is the documentation of what your infrastructure looks like.
Disaster recovery: Rebuild everything from code if a region goes down.
Cost visibility: Review infrastructure changes before they are applied (and before they start costing money).

Terraform vs other tools

Several IaC tools exist. Here’s how Terraform compares to the main alternatives:

Feature	Terraform	Pulumi	CloudFormation	Ansible
Language	HCL (declarative)	Python, TypeScript, Go, etc.	JSON/YAML	YAML (procedural)
Cloud support	Multi-cloud	Multi-cloud	AWS only	Multi-cloud (via modules)
State management	Explicit state file	Managed by Pulumi service	Managed by AWS	Stateless
Learning curve	Moderate	Varies by language	Moderate	Low
Ecosystem	Huge provider ecosystem	Growing	AWS-only but deep	Huge role ecosystem
Best for	Multi-cloud infra	Teams that prefer general-purpose languages	AWS-only shops	Configuration management

Terraform’s sweet spot is multi-cloud infrastructure provisioning with a declarative approach. If you’re on AWS only and want tight integration, CloudFormation is reasonable. If your team prefers writing Python over HCL, Pulumi deserves a look. But for most teams managing infrastructure across providers, Terraform is the pragmatic choice.

Core concepts

Providers

Providers are plugins that let Terraform interact with APIs — AWS, Azure, GCP, Kubernetes, GitHub, Cloudflare, and hundreds more.

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12


terraform {
 required_providers {
 aws = {
 source = "hashicorp/aws"
 version = "~> 5.0"
 }
 }
}

provider "aws" {
 region = "eu-west-1"
}

Resources

Resources are the fundamental building blocks. Each resource block describes one infrastructure object.

1
2
3
4
5
6
7
8


resource "aws_instance" "web" {
 ami = "ami-0c55b159cbfafe1f0"
 instance_type = "t3.micro"

 tags = {
 Name = "web-server"
 }
}

State

Terraform maintains a state file that maps your configuration to real-world resources. This is how Terraform knows what exists, what needs to change, and what to destroy. The state file is critical. Losing it means Terraform loses track of your infrastructure.

Modules

Modules are reusable packages of Terraform configuration. Think of them as functions: they take inputs (variables), create resources, and produce outputs.

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13


module "vpc" {
 source = "terraform-aws-modules/vpc/aws"
 version = "5.1.0"

 name = "my-vpc"
 cidr = "10.0.0.0/16"

 azs = ["eu-west-1a", "eu-west-1b"]
 private_subnets = ["10.0.1.0/24", "10.0.2.0/24"]
 public_subnets = ["10.0.101.0/24", "10.0.102.0/24"]

 enable_nat_gateway = true
}

Practical example: VPC + EC2

Here’s a complete example that provisions a VPC with a public subnet and an EC2 instance:

 1
 2
 3
 4
 5
 6
 7
 8
 9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114


terraform {
 required_version = ">= 1.5.0"
 required_providers {
 aws = {
 source = "hashicorp/aws"
 version = "~> 5.0"
 }
 }
}

provider "aws" {
 region = "eu-west-1"
}

# --- Networking ---

resource "aws_vpc" "main" {
 cidr_block = "10.0.0.0/16"
 enable_dns_support = true
 enable_dns_hostnames = true

 tags = {
 Name = "main-vpc"
 }
}

resource "aws_subnet" "public" {
 vpc_id = aws_vpc.main.id
 cidr_block = "10.0.1.0/24"
 availability_zone = "eu-west-1a"
 map_public_ip_on_launch = true

 tags = {
 Name = "public-subnet"
 }
}

resource "aws_internet_gateway" "gw" {
 vpc_id = aws_vpc.main.id

 tags = {
 Name = "main-igw"
 }
}

resource "aws_route_table" "public" {
 vpc_id = aws_vpc.main.id

 route {
 cidr_block = "0.0.0.0/0"
 gateway_id = aws_internet_gateway.gw.id
 }

 tags = {
 Name = "public-rt"
 }
}

resource "aws_route_table_association" "public" {
 subnet_id = aws_subnet.public.id
 route_table_id = aws_route_table.public.id
}

# --- Security Group ---

resource "aws_security_group" "web" {
 name = "web-sg"
 description = "Allow HTTP and SSH"
 vpc_id = aws_vpc.main.id

 ingress {
 from_port = 80
 to_port = 80
 protocol = "tcp"
 cidr_blocks = ["0.0.0.0/0"]
 }

 ingress {
 from_port = 22
 to_port = 22
 protocol = "tcp"
 cidr_blocks = ["YOUR_IP/32"] # Restrict to your IP
 }

 egress {
 from_port = 0
 to_port = 0
 protocol = "-1"
 cidr_blocks = ["0.0.0.0/0"]
 }
}

# --- EC2 Instance ---

resource "aws_instance" "web" {
 ami = "ami-0c55b159cbfafe1f0"
 instance_type = "t3.micro"
 subnet_id = aws_subnet.public.id
 vpc_security_group_ids = [aws_security_group.web.id]

 tags = {
 Name = "web-server"
 }
}

# --- Outputs ---

output "instance_public_ip" {
 value = aws_instance.web.public_ip
}

output "vpc_id" {
 value = aws_vpc.main.id
}

The plan/apply workflow

Terraform follows a predictable workflow:

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17


# 1. Initialize - download providers and modules
terraform init

# 2. Format - ensure consistent code style
terraform fmt

# 3. Validate - check syntax and configuration
terraform validate

# 4. Plan - preview what will change (critical step!)
terraform plan -out=tfplan

# 5. Apply - execute the plan
terraform apply tfplan

# 6. Destroy - tear down all resources (when needed)
terraform destroy

The terraform plan step is the most important. Never skip it. Always review the plan output before applying, especially in production. The plan shows you exactly what will be created, modified, or destroyed.

1
2


# Example plan output
Plan: 6 to add, 0 to change, 0 to destroy.

In CI/CD pipelines, save the plan to a file (-out=tfplan) and apply that exact plan. This prevents race conditions where infrastructure changes between the plan and apply steps.

State management best practices

State management is where most Terraform problems originate. Follow these practices:

Use a remote backend

Never store state locally or in Git. Use a remote backend with encryption and locking:

1
2
3
4
5
6
7
8
9


terraform {
 backend "s3" {
 bucket = "my-terraform-state"
 key = "prod/networking/terraform.tfstate"
 region = "eu-west-1"
 encrypt = true
 dynamodb_table = "terraform-locks"
 }
}

The DynamoDB table provides state locking. This prevents two people or pipelines from modifying the same infrastructure at the same time.

Organize state by component

Don’t put all your infrastructure in one state file. Split by component or team:

 1
 2
 3
 4
 5
 6
 7
 8
 9
10


environments/
├── prod/
│ ├── networking/ # VPC, subnets, routes
│ ├── compute/ # EC2, ASGs, load balancers
│ ├── database/ # RDS instances
│ └── monitoring/ # CloudWatch, alerts
└── staging/
 ├── networking/
 ├── compute/
 └── database/

Smaller state files mean faster plans, smaller blast radius, and fewer teams competing for locks.

Use `terraform_remote_state` sparingly

You can reference outputs from other state files, but use it carefully. Over-reliance on remote state creates tight coupling between components. Prefer passing values through variables or a parameter store.

Tips for production use

Pin provider versions. Use ~> constraints to allow patch updates but prevent breaking changes: version = "~> 5.0".
Use workspaces carefully. Workspaces are useful for simple environment separation but get confusing at scale. Separate directories per environment is usually clearer.
Implement a CI/CD pipeline for Terraform. Run terraform plan on PRs and post the output as a PR comment. Run terraform apply only after merge and approval.
Use prevent_destroy for critical resources. This lifecycle rule stops accidental destruction of databases or persistent storage:
1 2 3 4 5 6

resource "aws_db_instance" "main" { # ... lifecycle { prevent_destroy = true } }
Tag everything. Use a default_tags block in the provider to ensure every resource gets standard tags (environment, team, project).
Use tflint and checkov. Lint your Terraform code and scan for security misconfigurations before applying.
1 2 3

tflint --init tflint checkov -d .
Import existing resources. If you have manually created infrastructure, use terraform import to bring it under management instead of recreating it.
Review the plan diff carefully. A resource showing “destroy and recreate” might cause downtime. Understand which changes are in-place versus destructive.

Terraform is one of those tools that rewards discipline. The more consistently you follow these practices, the more confidently your team manages infrastructure at scale.

Hcl on Adur