Skip to main content
The terragrunt destroy command is your go-to solution for safely deprovisioning Terraform-managed infrastructure. By invoking this command, you can prevent unnecessary cloud costs, enforce clean-ups, and maintain consistent environments across your configurations.

​
Purpose

The image is a diagram titled "Terragrunt destroy," showing a central icon connected to two concepts: "Destruction or Resources" and "Safely removes infrastructure," with the word "Purpose" below.
Use terragrunt destroy when you need to remove resources provisioned by Terraform. It wraps Terraform’s native terraform destroy, extending it with Terragrunt’s configuration hierarchy and remote state management.

​
Workflow

The image is a workflow diagram for "Terragrunt destroy," showing steps like "Run after deployed and tested," "Deprovision resources," and "Release associated resources." It includes icons and a gradient "Workflow" label at the bottom.
A typical terragrunt destroy workflow involves:
  1. Deploying and validating infrastructure
  2. Executing the destroy command
  3. Deprovisioning resources
  4. Cleaning up state files and remote backends

​
Under the Hood

The image illustrates the process of "Terragrunt destroy," showing steps from initiating the command to actual resource destruction and its application across a hierarchy, with integration with Terraform.
When executed, Terragrunt traverses your configuration hierarchy, initializes each module, and delegates destruction to Terraform. This ensures consistent cleanup across all modules, environments, and remote states.

​
Confirmation Prompt

The image is about a "Terragrunt destroy" confirmation prompt, highlighting a feature that prompts users to avoid accidental deletions. It includes a speech bubble asking "Confirm deletion?" and a label "Confirmation Prompt."
Before any infrastructure is torn down, Terragrunt prompts you to confirm the destruction plan.
Always review the plan output before typing yes. Running terragrunt destroy is irreversible and may lead to data loss.

​
Parallel Execution

The image illustrates a process flow for "Terragrunt destroy" with a central icon connected to two puzzle piece icons, labeled "Parallel Execution."
Terragrunt can execute destroy operations in parallel across modules, drastically reducing total cleanup time for large, modular repositories.

​
Best Practices

EnvironmentRecommendation
Non-productionSchedule regular terragrunt destroy to reduce cost.
Development & StagingAutomate cleanup in CI/CD pipelines after testing.
Production & Always-onAvoid destructive commands; use targeted destroys.
  • Always run terragrunt plan before terragrunt destroy.
  • Lock remote state backends to prevent concurrent modifications.
  • Use --terragrunt-non-interactive in automated workflows.

​
Example: Destroying an AWS VPC

Below is a sample Terragrunt configuration that references an AWS VPC module. We’ll destroy the VPC and its associated resources: 
terraform {
  source = "tfr://terraform-aws-modules/vpc/aws/?version=5.8.1"
}

inputs = {
  name                          = "KodeKloud-VPC"
  vpc_cidr_block                = "10.0.0.0/16"
  vpc_enable_dns_hostnames      = true
  vpc_enable_dns_support        = true
  vpc_instance_tenancy          = "default"
  vpc_flow_log_destination_type = "cloud-watch-logs"

  # Identifiers for existing resources to be destroyed
  vpc_id                  = "vpc-03218326bbd74f45c"
  vpc_arn                 = "arn:aws:ec2:us-east-1:654654587809:vpc/vpc-03218326bbd74f45c"
  vpc_main_route_table_id = "rtb-0f620361b604057a9"
  vpc_owner_id            = "654654587809"

  # Empty lists ensure Terraform will remove these associations
  private_route_table_ids     = []
  public_subnets              = []
  private_subnets_cidr_blocks = tolist([])
  public_subnets_cidr_blocks  = tolist([])
}
Run the destroy command in your VPC module directory: 
cd path/to/vpc
terragrunt destroy
Inspect the plan to confirm the intended resources are queued for removal: 
Plan: 0 to add, 0 to change, 4 to destroy.

Changes to Outputs:
  - azs                           = "null" -> null
  - default_route_table_id        = "acl-00cf3b95846d6be5" -> null
  - cgw_arns                      = [] -> null
  - cgw_ids                       = [] -> null
  # (additional output removals)
Type yes when prompted to proceed. Terragrunt will then deprovision the VPC and all related resources.

Watch Video