> ## Documentation Index
> Fetch the complete documentation index at: https://enterprise-docs.crewai.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Terraform Infrastructure on AWS

> Provision AWS infrastructure for CrewAI Platform with Terraform and map outputs to Helm values.

## Overview

Use this guide to provision baseline AWS infrastructure for CrewAI Platform with Terraform, then continue with Helm installation.

<Warning>
  Request the Terraform files from the CrewAI Factory Engineering team before following this guide.
</Warning>

This flow aligns with the AWS Terraform example in the installer and covers:

* Amazon EKS cluster
* Aurora PostgreSQL
* S3 buckets for platform data and logs
* ECR repository for crew image builds
* Cluster add-ons required by the platform

## Prerequisites

* Terraform `>= 1.13.4`
* AWS CLI configured for the target account and region
* `kubectl` and Helm installed
* AWS permissions to create networking, EKS, RDS, IAM, S3, ECR, and Secrets Manager resources

## Step 1: Configure Terraform State

From your `crewai-installer` checkout, open `terraform/examples/aws/main.tf`.

The example uses an S3 backend for remote state:

```hcl theme={null}
backend "s3" {
  bucket  = "YOUR-UNIQUE-STATE-BUCKET-NAME"
  key     = "crewai/terraform.tfstate"
  region  = "us-east-1"
  encrypt = true
}
```

* Create the state bucket before running `terraform init`.
* For temporary testing only, you can comment out the backend block and use local state.

## Step 2: Configure Variables

```bash theme={null}
cd terraform/examples/aws
cp terraform.tfvars.example terraform.tfvars
```

Update at least:

* `region`
* `environment`
* `s3_data_bucket_name` (globally unique)
* `s3_logs_bucket_name` (globally unique)

The example also auto-detects:

* The IAM identity running Terraform
* The public IP running Terraform (added to cluster access CIDRs)

You can still provide additional values with:

* `admin_iam_principals`
* `cluster_endpoint_public_access_cidrs`

## Step 3: Apply Infrastructure

```bash theme={null}
terraform init
terraform plan
terraform apply
```

When apply finishes, configure `kubectl` using the Terraform output command:

```bash theme={null}
terraform output configure_kubectl_command
```

## Step 4: Map Terraform Outputs to Helm Values

Use Terraform outputs to populate your `my-values.yaml` for chart install:

* `database_endpoint` -> `envVars.DB_HOST`
* `s3_data_bucket` -> `envVars.AWS_BUCKET`
* `ecr_repository_url` -> `envVars.CREW_IMAGE_REGISTRY_OVERRIDE`

You still need to set application-specific values such as:

* `envVars.APPLICATION_HOST`
* ingress class and TLS settings
* authentication provider settings

Continue with the [Installation Guide](/installation/installation).

## Step 5: Destroy Safely (When Needed)

Before `terraform destroy`, delete Kubernetes Ingress resources managed by AWS Load Balancer Controller.

If Ingress resources remain, destroy can hang while AWS load balancer dependencies are still attached.

## Troubleshooting Notes

* If `terraform init` fails, confirm your backend bucket exists and backend values are correct.
* If `kubectl` cannot reach the cluster API, verify your public IP is authorized.
* If chart deployment cannot push crew images, verify `envVars.CREW_IMAGE_REGISTRY_OVERRIDE` points to a writable ECR registry.

## Next Steps

* [Installation Guide](/installation/installation)
* [Configuration Guide](/configuration/configuration)
* [AWS Integration Guide](/cloud-providers/aws)