Terraform Generation & Deployment
AIVory Architect generates production-ready Terraform code from your visual designs and can deploy infrastructure directly from your IDE.
- Production-ready code generated from visual designs
- Dependency resolution automatically orders resources correctly
- Diff viewer shows changes before overwriting files
- One-click deployment directly from your IDE (Premium)
- State tracking shows deployed status on canvas
Code Generation
Generated Files
When you click Generate Terraform, Architect creates:
| File | Contents |
|---|---|
main.tf |
All resource definitions |
variables.tf |
Input variables for configuration |
outputs.tf |
Output values (IPs, endpoints, etc.) |
providers.tf |
Provider configuration blocks |
terraform.tfvars.example |
Example variable values |
Files are saved to .aivory/terraform/ by default.
Generation Process
Resource Collection
All canvas resources are collected
Dependency Resolution
Connections are analyzed for ordering
Topological Sort
Resources are ordered by dependencies
Template Rendering
Provider-specific templates generate HCL
Variable Extraction
Secrets and config become variables
Output Generation
Resource attributes become outputs
Diff Viewer
Before overwriting existing files, Architect shows a diff view:
- Green lines: New additions
- Red lines: Removed content
- Yellow lines: Modified content
Note
Review changes before confirming to prevent accidental overwrites.
Generated Code Example
Visual Design
Generated Terraform
View full generated code example
# providers.tf
terraform {
required_providers {
hcloud = {
source = "hetznercloud/hcloud"
version = "~> 1.45"
}
}
}
provider "hcloud" {
token = var.hcloud_token
}
# variables.tf
variable "hcloud_token" {
description = "Hetzner Cloud API token"
type = string
sensitive = true
}
variable "server_type" {
description = "Server type for web server"
type = string
default = "cx21"
}
# main.tf
resource "hcloud_network" "main" {
name = "main-network"
ip_range = "10.0.0.0/8"
}
resource "hcloud_network_subnet" "main" {
network_id = hcloud_network.main.id
type = "cloud"
network_zone = "eu-central"
ip_range = "10.0.1.0/24"
}
resource "hcloud_server" "web" {
name = "web-server"
server_type = var.server_type
image = "ubuntu-22.04"
location = "nbg1"
network {
network_id = hcloud_network.main.id
ip = "10.0.1.2"
}
depends_on = [hcloud_network_subnet.main]
}
resource "hcloud_volume" "data" {
name = "data-volume"
size = 50
location = "nbg1"
}
resource "hcloud_volume_attachment" "data" {
volume_id = hcloud_volume.data.id
server_id = hcloud_server.web.id
automount = true
}
# outputs.tf
output "server_ip" {
description = "Public IP of the web server"
value = hcloud_server.web.ipv4_address
}
output "server_private_ip" {
description = "Private IP of the web server"
value = hcloud_server.web.network[*].ip
}
Deployment (Premium)
Success
Direct deployment from your IDE requires a premium subscription.
Deployment Workflow
1. Click Deploy
Opens the deployment panel
2. Terraform Init
Automatically initializes providers
3. Terraform Plan
Shows what will be created/modified/destroyed
4. Review
Examine the plan for unexpected changes
5. Confirm
Click to run
terraform apply
6. Monitor
Watch real-time progress and logs
Deployment Panel
The deployment panel shows:
- Status: Current operation (init/plan/apply)
- Progress: Visual progress indicator
- Logs: Streaming Terraform output
- Resources: List of resources being managed
- Outputs: Values from
outputs.tfafter apply
Cancel Operations
Long-running operations can be cancelled:
- Click Cancel in the deployment panel
- Terraform gracefully stops
- Partial changes may remain - check state
Warning
Cancelling during apply may leave resources in an inconsistent state. Always verify with
terraform plan after cancellation.
State Management
State File
Terraform tracks deployed resources in terraform.tfstate:
- Architect reads this file to show deployed status
- Never edit the state file manually
- Back up state files before major changes
Note
For team collaboration, use a remote backend (S3, Azure Blob, etc.) instead of local state files.
Deployed Resource Tracking
After deployment, the canvas shows:
Green Indicator
Resource deployed successfully
Yellow Indicator
Resource has drift (manual changes detected)
Red Indicator
Resource failed to deploy
Refresh State
Click Refresh to update the canvas from current state:
- Reads
terraform.tfstate - Updates resource status indicators
- Detects any drift from design
Advanced Configuration
Custom Backend
Add backend configuration for remote state:
# backend.tf (create manually)
terraform {
backend "s3" {
bucket = "my-terraform-state"
key = "project/terraform.tfstate"
region = "us-east-1"
}
}
Variable Files
Create terraform.tfvars for environment-specific values:
hcloud_token = "your-token-here"
server_type = "cx31"
Or use separate files per environment:
terraform apply -var-file="production.tfvars"
Workspaces
Terraform workspaces let you manage multiple environments:
terraform workspace new staging
terraform workspace select staging
terraform apply
Troubleshooting
Provider not found
Run
terraform init manually or click Initialize in Architect.
Resource already exists
The resource exists but isn’t in state. Either:
- Import it:
terraform import resource_type.name id - Delete it manually and re-apply
Cycle detected
Circular dependency in your design. Review connections and break the cycle.
Permission denied
Check your cloud provider credentials in
.env or credential vault.
Best Practices
Version Control
Commit
.tf files to git. Don’t commit state or credentials.
Code Review
Review generated code before deploying. Use
terraform plan to preview.
Modular Design
Group related resources. Use consistent naming conventions.
Test First
Test in a staging environment before deploying to production.
Note
Consider splitting large designs into modules for better maintainability and reusability.