Terraform Resource

Terraform hub

A Terraform resource is a block of HCL configuration that declares one piece of infrastructure — like an EC2 instance, an S3 bucket, or a DNS record. Terraform creates, updates, and deletes the real-world object to match the resource declaration.

What Is a Terraform Resource?

A resource is the most important block in Terraform. Every piece of infrastructure you want to exist — an EC2 instance, an S3 bucket, a Route 53 record, a GitHub repository — is declared as a resource block. Terraform's entire job is to make the real world match what your resource blocks say.

Think of a resource block as a specification: "I want an S3 bucket named swiggy-order-images-prod with versioning enabled and public access blocked." Terraform reads that spec, compares it to what actually exists, and makes the minimum changes needed.

Resource Block Anatomy

HCL

# resource "<TYPE>" "<NAME>" — both parts together form the address
resource "aws_s3_bucket" "order_images" {
  # Arguments — specific to this resource type
  bucket = "swiggy-order-images-prod"   # the S3 bucket name
 
  tags = {
    Environment = "production"
    Team        = "platform"
  }
}

The resource address aws_s3_bucket.order_images is how you reference this resource from other parts of your configuration.

Referencing Resource Attributes

After Terraform creates a resource, you can read its attributes in other resource blocks:

HCL

resource "aws_s3_bucket" "order_images" {
  bucket = "swiggy-order-images-prod"
}
 
resource "aws_s3_bucket_versioning" "order_images" {
  bucket = aws_s3_bucket.order_images.id   # reads the bucket ID after creation
 
  versioning_configuration {
    status = "Enabled"
  }
}

This reference also creates an implicit dependency — Terraform knows to create the bucket before the versioning config.

Meta-Arguments

Meta-arguments are special arguments that work on any resource type, regardless of provider:

HCL

resource "aws_instance" "app_server" {
  ami           = "ami-0f5ee92e2d63afc18"
  instance_type = "t3.medium"
 
  # count — create multiple identical instances
  count = 3   # creates app_server[0], app_server[1], app_server[2]
 
  # depends_on — explicit dependency when there is no attribute reference
  depends_on = [aws_iam_instance_profile.app_profile]
 
  # lifecycle — control replace/destroy behaviour
  lifecycle {
    create_before_destroy = true   # spin up replacement before killing old one
    prevent_destroy       = true   # block terraform destroy for this resource
  }
 
  # provider — use a non-default provider
  provider = aws.us_east
 
  tags = { Name = "app-server-${count.index}" }
}

for_each — Creating Resources from a Map

HCL

# Create one S3 bucket per environment
resource "aws_s3_bucket" "env_buckets" {
  for_each = toset(["dev", "staging", "prod"])   # set of strings
 
  bucket = "zerodha-data-${each.key}"   # each.key = dev / staging / prod
 
  tags = { Environment = each.key }
}
# Creates: aws_s3_bucket.env_buckets["dev"]
#          aws_s3_bucket.env_buckets["staging"]
#          aws_s3_bucket.env_buckets["prod"]

Lifecycle Rules

HCL

resource "aws_db_instance" "payments_db" {
  identifier     = "payments-prod"
  engine         = "postgres"
  instance_class = "db.t3.medium"
  username       = "admin"
  password       = var.db_password
 
  lifecycle {
    prevent_destroy       = true    # block accidental terraform destroy
    create_before_destroy = true    # for zero-downtime replacement
    ignore_changes        = [password]  # don't revert manual password rotations
  }
}

Troubleshooting Resources

Error	Cause	Fix
`Error: Reference to undeclared resource`	Typo in resource address	Check `resource_type.resource_name` spelling
`Error: Unsupported argument`	Wrong argument name for this resource type	Check provider documentation for valid arguments
`Error: Resource already exists`	Resource exists in AWS but not in state	Use `terraform import` to bring it into state
`Error: timeout while waiting for resource`	AWS API slow or limits hit	Increase timeout in `timeouts` block

Tip
Use terraform state list to see every resource Terraform is tracking. Use terraform state show aws_s3_bucket.order_images to inspect all attributes of a specific resource.

Common Mistake
Using count when you should use for_each. If you have count = 3 and remove the first item, Terraform renumbers all resources and may destroy and recreate resources you did not intend to touch. Use for_each with a map for any resource you might add or remove from.

Frequently Asked Questions

What's the relationship between a Terraform resource block and the real cloud object it manages?

Each resource block is tracked in the state file by a unique resource address (like `aws_instance.web`), which is mapped internally to the cloud provider's real resource ID (like an EC2 instance ID) — Terraform uses that mapping to know what to check, update, or destroy on each run. Deleting a resource block from your .tf files and running apply will destroy the real object; removing it from state with `terraform state rm` instead just stops Terraform from tracking it, leaving the real resource untouched.

What's a common mistake that causes an unwanted resource replacement instead of an in-place update?

Changing an attribute that a provider marks as `ForceNew` in its schema — such as an EC2 instance's availability zone or an RDS engine version in some configurations — triggers a destroy-and-recreate instead of an in-place update, which is often not what the engineer expects or wants for a stateful resource. Always check `terraform plan` output for `-/+` (replace) versus `~` (update in place) before applying, especially on anything holding data.