Day 8: Terraform Finally Makes Sense (Loops, Count, For_Each & Dependencies)

Source: Dev.to

Overview

Today was one of those days where everything just clicked. I finally understood Terraform meta‑arguments—especially count, for_each, and depends_on. After some hands‑on practice, the concepts that once seemed scary turned out to be simple and powerful.

Using count

count is great for iterating over lists. In the example below, a list of bucket names is supplied via a variable, and count.index is used to create two S3 buckets:

resource "aws_s3_bucket" "bucket5" {
  count  = 2
  bucket = var.bucket_name[count.index]
  tags   = var.tags
}

Running the typical Terraform workflow:

terraform plan
terraform apply --auto-approve
terraform output

produced the expected resources without any drama.

Using for_each

When you need unique resource names, for_each works cleaner with sets. The following snippet demonstrates creating two more S3 buckets using a set and referencing each bucket name with each.value:

resource "aws_s3_bucket" "bucket_set" {
  for_each = var.bucket_set   # a set of bucket names
  bucket   = each.value
  tags     = var.tags
}

Managing Dependencies with depends_on

Explicit dependencies ensure Terraform doesn’t rush resources in the wrong order. Adding depends_on guarantees that a resource waits for another to be fully provisioned before proceeding:

resource "aws_s3_bucket" "dependent_bucket" {
  bucket = "my-dependent-bucket"
  depends_on = [
    aws_s3_bucket.bucket5
  ]
}

Clean Outputs with For‑Expressions

For‑expressions make output values concise and readable. The example below collects the names of all buckets created with count:

output "s3_bucket_names" {
  value = [for bucket in aws_s3_bucket.bucket5 : bucket.bucket]
}

The output is a tidy list of bucket names, suitable for downstream consumption.

Takeaways

• count – ideal for iterating over ordered lists.
• for_each – cleaner when you need unique identifiers, especially with sets.
• depends_on – prevents Terraform from applying resources out of order.
• For‑expressions – produce neat, professional outputs.

The more I practice Terraform, the more confident I feel about doing DevOps work. “Okay… I can actually do this DevOps thing.”