Skip to main content
The Terragrunt dependency block lets you declare dependencies between modules, enabling one module to consume another’s outputs. This promotes modular infrastructure, improves code reuse, and streamlines inter-module communication.
Using dependency blocks helps enforce a clear execution order and guarantees that required resources are available before they’re referenced.

​
Key Attributes

AttributeType / DefaultDescription
namestringIdentifier for the dependency block. Useful when you have multiple dependencies.
config_pathstringPath to the Terragrunt module or folder with the terragrunt.hcl.
enabledboolean / trueControls whether Terragrunt fetches this dependency. Set to false to skip resolution.
skip_outputsboolean / falseIf true, skips executing terragrunt output, reducing runtime when outputs aren’t needed.
mock_outputsmapKey-value pairs to return when the target module has no real outputs.
mock_outputs_allowedlistTerraform commands (e.g., plan, apply) for which mock outputs can be used.
mock_outputs_merge_strategy_with_statestringStrategy for merging mock outputs with existing remote state (e.g., merge, overwrite).

​
Benefits of Using Dependency Blocks

  • Modularization & Reusability
    Break your infrastructure into discrete modules that depend on each other’s outputs.
  • Output References
    Access values via dependency.<name>.outputs.<key> expressions.
  • Selective Execution
    Temporarily disable dependencies using the enabled flag.
  • Mock Outputs for Testing
    Simulate module outputs during validate or plan without applying the dependent module.
  • Performance Optimization
    Skip unnecessary state downloads with skip_outputs and let Terragrunt fetch only the required outputs.
  • Team Collaboration
    Independent development and testing of modules with clearly defined input/output contracts.

​
Optimization Strategies

Terragrunt fetches dependencies in parallel at each root level but parses nested dependencies serially. When using remote state and optimization flags, it retrieves only the lowest-level outputs, significantly speeding up large infrastructures.
Overusing mock outputs can mask real configuration issues. Always validate with real outputs before production deployments.

​
Best Practices

  • Organize dependencies in a dedicated block at the top of your terragrunt.hcl.
  • Keep config_path references relative and consistent.
  • Use mock_outputs only for CI/CD testing and validation.
  • Ensure remote state backends and optimization flags (skip_outputs) are configured for faster runs.
  • Name your dependencies meaningfully to convey the resource relationship.

​
Example Usage

​
1. EC2 Module Without Dependencies

terraform {
  source = "tfr://terraform-aws-modules/ec2-instance/aws//?version=5.6.1"
}

include "root" {
  path   = find_in_parent_folders()
  expose = true
}

include "common" {
  path   = find_in_parent_folders("common.hcl")
  expose = true
}

inputs = {
  name = include.common.locals.project
}

​
2. Adding a VPC Dependency

By adding a dependency block, you ensure the EC2 instance uses the VPC created by another module: 
terraform {
  source = "tfr://terraform-aws-modules/ec2-instance/aws//?version=5.6.1"
}

include "root" {
  path   = find_in_parent_folders()
  expose = true
}

include "common" {
  path   = find_in_parent_folders("common.hcl")
  expose = true
}

dependency "vpc" {
  config_path = "../vpc"
}

inputs = {
  name      = include.common.locals.project
  subnet_id = dependency.vpc.outputs.public_subnets[0]
}
This configuration guarantees that the VPC is provisioned before the EC2 instance and that the instance launches in the correct public subnet.

Watch Video