Skip to main content

Using Remote State

Terraform natively supports the concept of remote state and there's a very easy way to access the outputs of one Terraform component in another component. We simplify this using the remote-state module, which is stack-aware and can be used to access the remote state of a component in the same or a different Atmos stack.

As your architecture grows, it helps to be more intentional about how you deconstruct and organize your components to keep your Terraform state small (see our best practices). By creating smaller components, your state becomes naturally more manageable. However, this introduces a new problem: there are now dependencies between your components, and the state becomes distributed. We need to find a new way for state to flow between your components and for and a way to share configuration. Plus, we want to avoid manual duplication of configurations as much as possible because that leads to bugs, like copy-paste mistakes.

You will learn

  • How to use the remote-state module to access the remote state of a component in the same or a different Atmos stack
  • How to configure Atmos to work the remote-state module to access the remote state of a component
  • Alternatives that might be easier for your use case

In Atmos, this is solved by using these modules:

  • terraform-provider-utils - The Cloud Posse Terraform Provider for various utilities, including stack configuration management

  • remote-state - Terraform module that loads and processes stack configurations from YAML sources and returns remote state outputs for Terraform components

The terraform-provider-utils is implemented in Go and uses Atmos Go modules to work with Atmos CLI config and Atmos stacks. The provider processes stack configurations to get the final config for an Atmos component in an Atmos stack. The final component config is then used by the remote-state Terraform module to return the remote state for the component in the stack.

NOTE:

Terraform remote state is incompatible with the local backend type. This is because the local backend is not recommended for production. Review the alternatives here, or consider switching to one of the other backend types.

New & Improved Way to Share Data

Atmos now supports a new way to share data between components using a template function in your Stack configurations:

{{ atmos.Component("<something>").outputs.<output_name> }}

This function allows reading any Atmos section or any attribute (not just outputs) from a section of an Atmos component in a stack. For more details, refer to the atmos.Component function.

Example

Here is an example.

Suppose that we need to provision two Terraform components:

The vpc Terraform component needs the outputs from the vpc-flow-logs-bucket Terraform component to configure VPC Flow Logs and store them in the S3 bucket.

We will provision the two Terraform components in the ue2-dev Atmos stack (in the dev AWS account by setting stage = "dev" and in the us-east-2 region by setting environment = "ue2").

Configure and Provision the vpc-flow-logs-bucket Component

In the stacks/catalog/vpc-flow-logs-bucket.yaml file, add the following default configuration for the vpc-flow-logs-bucket/defaults Atmos component:

stacks/catalog/vpc-flow-logs-bucket.yaml

components:
terraform:
vpc-flow-logs-bucket/defaults:
metadata:
# `metadata.type: abstract` makes the component `abstract`,
# explicitly prohibiting the component from being deployed.
# `atmos terraform apply` will fail with an error.
# If `metadata.type` attribute is not specified, it defaults to `real`.
# `real` components can be provisioned by `atmos` and CI/CD like Spacelift and Atlantis.
type: abstract
# Default variables, which will be inherited and can be overridden in the derived components
vars:
force_destroy: false
lifecycle_rule_enabled: false
traffic_type: "ALL"

In the stacks/ue2-dev.yaml stack config file, add the following config for the vpc-flow-logs-bucket-1 Atmos component in the ue2-dev Atmos stack:

stacks/ue2-dev.yaml

# Import the base Atmos component configuration from the `catalog`.
# `import` supports POSIX-style Globs for file names/paths (double-star `**` is supported).
# File extensions are optional (if not specified, `.yaml` is used by default).
import:
- catalog/vpc-flow-logs-bucket

components:
terraform:
vpc-flow-logs-bucket-1:
metadata:
# Point to the Terraform component in `components/terraform` folder
component: infra/vpc-flow-logs-bucket
inherits:
# Inherit all settings and variables from the
# `vpc-flow-logs-bucket/defaults` base Atmos component
- vpc-flow-logs-bucket/defaults
vars:
# Define variables that are specific for this component
# and are not set in the base component
name: vpc-flow-logs-bucket-1
# Override the default variables from the base component
traffic_type: "REJECT"

Having the stacks configured as shown above, we can now provision the vpc-flow-logs-bucket-1 Atmos component into the ue2-dev stack by executing the following Atmos commands:

atmos terraform plan vpc-flow-logs-bucket-1 -s ue2-dev
atmos terraform apply vpc-flow-logs-bucket-1 -s ue2-dev

Configure and Provision the vpc Component

Having the vpc-flow-logs-bucket Terraform component provisioned into the ue2-dev stack, we can now configure the vpc Terraform component to obtain the outputs from the remote state of the vpc-flow-logs-bucket-1 Atmos component.

In the components/terraform/infra/vpc/remote-state.tf file, configure the remote-state Terraform module to obtain the remote state for the vpc-flow-logs-bucket-1 Atmos component:

components/terraform/infra/vpc/remote-state.tf

module "vpc_flow_logs_bucket" {
count = local.vpc_flow_logs_enabled ? 1 : 0

source = "cloudposse/stack-config/yaml//modules/remote-state"
version = "1.5.0"

# Specify the Atmos component name (defined in YAML stack config files)
# for which to get the remote state outputs
component = var.vpc_flow_logs_bucket_component_name

# Override the context variables to point to a different Atmos stack if the
# `vpc-flow-logs-bucket-1` Atmos component is provisioned in another AWS account, OU or region
stage = try(coalesce(var.vpc_flow_logs_bucket_stage_name, module.this.stage), null)
tenant = try(coalesce(var.vpc_flow_logs_bucket_tenant_name, module.this.tenant), null)
environment = try(coalesce(var.vpc_flow_logs_bucket_environment_name, module.this.environment), null)

# `context` input is a way to provide the information about the stack (using the context
# variables `namespace`, `tenant`, `environment`, `stage` defined in the stack config)
context = module.this.context
}

In the components/terraform/infra/vpc/vpc-flow-logs.tf file, configure the aws_flow_log resource for the vpc Terraform component to use the remote state output vpc_flow_logs_bucket_arn from the vpc-flow-logs-bucket-1 Atmos component:

components/terraform/infra/vpc/vpc-flow-logs.tf

locals {
enabled = module.this.enabled
vpc_flow_logs_enabled = local.enabled && var.vpc_flow_logs_enabled
}

resource "aws_flow_log" "default" {
count = local.vpc_flow_logs_enabled ? 1 : 0

# Use the remote state output `vpc_flow_logs_bucket_arn` of the `vpc_flow_logs_bucket` component
log_destination = module.vpc_flow_logs_bucket[0].outputs.vpc_flow_logs_bucket_arn

log_destination_type = var.vpc_flow_logs_log_destination_type
traffic_type = var.vpc_flow_logs_traffic_type
vpc_id = module.vpc.vpc_id

tags = module.this.tags
}

In the stacks/catalog/vpc.yaml file, add the following default config for the vpc/defaults Atmos component:

stacks/catalog/vpc.yaml

components:
terraform:
vpc/defaults:
metadata:
# `metadata.type: abstract` makes the component `abstract`,
# explicitly prohibiting the component from being deployed.
# `atmos terraform apply` will fail with an error.
# If `metadata.type` attribute is not specified, it defaults to `real`.
# `real` components can be provisioned by `atmos` and CI/CD like Spacelift and Atlantis.
type: abstract
# Default variables, which will be inherited and can be overridden in the derived components
vars:
public_subnets_enabled: false
nat_gateway_enabled: false
nat_instance_enabled: false
max_subnet_count: 3
vpc_flow_logs_enabled: false
vpc_flow_logs_log_destination_type: s3
vpc_flow_logs_traffic_type: "ALL"

In the stacks/ue2-dev.yaml stack config file, add the following config for the vpc/1 Atmos component in the ue2-dev stack:

stacks/ue2-dev.yaml

# Import the base component configuration from the `catalog`.
# `import` supports POSIX-style Globs for file names/paths (double-star `**` is supported).
# File extensions are optional (if not specified, `.yaml` is used by default).
import:
- catalog/vpc

components:
terraform:
vpc/1:
metadata:
# Point to the Terraform component in `components/terraform` folder
component: infra/vpc
inherits:
# Inherit all settings and variables from the `vpc/defaults` base Atmos component
- vpc/defaults
vars:
# Define variables that are specific for this component
# and are not set in the base component
name: vpc-1
ipv4_primary_cidr_block: 10.8.0.0/18
# Override the default variables from the base component
vpc_flow_logs_enabled: true
vpc_flow_logs_traffic_type: "REJECT"

# Specify the name of the Atmos component that provides configuration
# for the `infra/vpc-flow-logs-bucket` Terraform component
vpc_flow_logs_bucket_component_name: vpc-flow-logs-bucket-1

# Override the context variables to point to a different Atmos stack if the
# `vpc-flow-logs-bucket-1` Atmos component is provisioned in another AWS account, OU or region.

# If the bucket is provisioned in a different AWS account,
# set `vpc_flow_logs_bucket_stage_name`
# vpc_flow_logs_bucket_stage_name: prod

# If the bucket is provisioned in a different AWS OU,
# set `vpc_flow_logs_bucket_tenant_name`
# vpc_flow_logs_bucket_tenant_name: core

# If the bucket is provisioned in a different AWS region,
# set `vpc_flow_logs_bucket_environment_name`
# vpc_flow_logs_bucket_environment_name: uw2

Having the stacks configured as shown above, we can now provision the vpc/1 Atmos component into the ue2-dev stack by executing the following Atmos commands:

atmos terraform plan vpc/1 -s ue2-dev
atmos terraform apply vpc/1 -s ue2-dev

Atmos Configuration

Both the atmos CLI and terraform-provider-utils Terraform provider use the same Go code, which try to locate the CLI config atmos.yaml file before parsing and processing Atmos stacks.

This means that atmos.yaml file must be at a location in the file system where all the processes can find it.

While placing atmos.yaml at the root of the repository will work for Atmos, it will not work for the terraform-provider-utils Terraform provider because the provider gets executed from the component's directory (e.g. components/terraform/infra/vpc), and we don't want to replicate atmos.yaml into every component's folder.

info

atmos.yaml is loaded from the following locations (from lowest to highest priority):

  • System dir (/usr/local/etc/atmos/atmos.yaml on Linux, %LOCALAPPDATA%/atmos/atmos.yaml on Windows)
  • Home dir (~/.atmos/atmos.yaml)
  • Current directory
  • ENV variables ATMOS_CLI_CONFIG_PATH and ATMOS_BASE_PATH
NOTE:

Initial Atmos configuration can be controlled by these ENV vars:

  • ATMOS_CLI_CONFIG_PATH - where to find atmos.yaml. Absolute path to a folder where the atmos.yaml CLI config file is located
  • ATMOS_BASE_PATH - absolute path to the folder containing the components and stacks folders

For this to work for both the atmos CLI and the Terraform provider, we recommend doing one of the following:

  • Put atmos.yaml at /usr/local/etc/atmos/atmos.yaml on local host and set the ENV var ATMOS_BASE_PATH to point to the absolute path of the root of the repo

  • Put atmos.yaml into the home directory (~/.atmos/atmos.yaml) and set the ENV var ATMOS_BASE_PATH pointing to the absolute path of the root of the repo

  • Put atmos.yaml at a location in the file system and then set the ENV var ATMOS_CLI_CONFIG_PATH to point to that location. The ENV var must point to a folder without the atmos.yaml file name. For example, if atmos.yaml is at /atmos/config/atmos.yaml, set ATMOS_CLI_CONFIG_PATH=/atmos/config. Then set the ENV var ATMOS_BASE_PATH pointing to the absolute path of the root of the repo

  • When working in a Docker container, place atmos.yaml in the rootfs directory at /rootfs/usr/local/etc/atmos/atmos.yaml and then copy it into the container's file system in the Dockerfile by executing the COPY rootfs/ / Docker command. Then in the Dockerfile, set the ENV var ATMOS_BASE_PATH pointing to the absolute path of the root of the repo. Note that the Atmos example uses Geodesic as the base Docker image. Geodesic sets the ENV var ATMOS_BASE_PATH automatically to the absolute path of the root of the repo on local host

Summary

  • Remote State for an Atmos component in an Atmos stack is obtained by using the remote-state Terraform module

  • The module calls the terraform-provider-utils Terraform provider which processes the stack configs and returns the configuration for the Atmos component in the stack. The terraform-provider-utils Terraform provider utilizes Atmos Go modules to parse and process stack configurations

  • The remote-state module accepts the component input as the Atmos component name for which to get the remote state outputs

  • The module accepts the context input as a way to provide the information about the stack (using the context variables namespace, tenant, environment, stage defined in the stack manifests)

  • If the Atmos component (for which we want to get the remote state outputs) is provisioned in a different Atmos stack (in a different AWS OU, or different AWS account, or different AWS region), we can override the context variables tenant, stage and environment to point the module to the correct stack. For example, if the component is provisioned in a different AWS region (let's say us-west-2), we can set environment = "uw2", and the remote-state module will get the remote state outputs for the Atmos component provisioned in that region

Use Remote State in Stack Configurations

Atmos supports an alternative way to read the outputs (remote state) of components directly in Atmos stack manifests by using the atmos.Component Go template function and the outputs section instead of using the remote-state module and configuring Terraform/OpenTofu components to use the module.

Learn How