Temporal Terraform Orchestrator
A Temporal-based workflow orchestration system for managing multi-workspace Terraform deployments with dependency resolution, variable passing between workspaces, and MCP server integration for AI-driven automation.
Potencialize seu agente
Instalação do publicador. Não é um endpoint do PlugMCP.
Host a copy with our keys and logs.
Ainda não há uma versão empacotada no registro. Siga o README do repositório.
URL remota ou um comando na sua máquina? Como as duas instalações diferem
Rodar no PlugMCP
A new server you create. It is not this listing's official endpoint. Use the publisher install when you want their URL and their credentials.
AbrirDo publicador
A Temporal-based workflow orchestration system for managing multi-workspace Terraform deployments with dependency resolution, variable passing between workspaces, and MCP server integration for AI-driven automation.
A Temporal-based workflow orchestration system for managing multi-workspace Terraform deployments with dependency resolution, variable passing between workspaces, and MCP server integration for AI-driven automation.
This project provides a robust infrastructure-as-code orchestration layer using Temporal workflows to manage Terraform operations across multiple workspaces with:
ParentWorkflow (Orchestrator)
│
├─> TerraformWorkflow (vpc) ← runs first (no dependencies)
│
├─> TerraformWorkflow (vpc-2) ← runs in parallel with vpc
│
├─> TerraformWorkflow (subnets) ← waits for vpc, receives vpc_id output
│
└─> TerraformWorkflow (eks) ← waits for vpc + subnets, receives both outputs
ParentWorkflow: The orchestrator that:
TerraformWorkflow: Executes Terraform operations for a single workspace:
terraform init - Initialize the workspaceterraform plan - Create execution plan (with -detailed-exitcode to detect changes)terraform show -json - Validate the plan outputterraform apply - Apply changes (skipped if no changes detected)terraform output -json - Capture outputs for downstream workspacesChild workflows are spawned as nested children of their "host" workflow (the deepest dependency). This creates a natural hierarchy where:
localhost:7233 (or set TEMPORAL_ADDRESS)Install dependencies:
go mod tidy
Start Temporal (if not already running):
temporal server start-dev
Start the worker:
go run ./cmd/worker
Execute the workflow:
go run ./cmd/starter -config infra.yaml
Monitor execution in the Temporal Web UI at http://localhost:8233
The CLI starter (cmd/starter) initiates workflow execution from a YAML configuration file.
go run ./cmd/starter [flags]
# Run with default settings
go run ./cmd/starter
# Specify a custom config file
go run ./cmd/starter -config ./environments/production.yaml
# Use a custom workflow ID for tracking
go run ./cmd/starter -config infra.yaml -workflow-id "deploy-prod-2024-01-15"
workspace_rootThe MCP (Model Context Protocol) server enables AI agents and automation tools to interact with the orchestration system.
go run ./cmd/mcp-server
The server runs on stdio and communicates via JSON-RPC, following the MCP specification.
list_workflowsLists available workflows and configured workspaces from the configuration file.
Parameters:
Response example:
{
"config_path": "infra.yaml",
"workspace_root": ".",
"workflows": [
{
"name": "ParentWorkflow",
"description": "Orchestrates terraform operations across multiple workspaces with dependencies",
"configured_workspaces": [
{
"name": "vpc",
"kind": "terraform",
"dir": "/abs/path/vpc",
"dependsOn": []
},
{
"name": "subnets",
"kind": "terraform",
"dir": "/abs/path/subnets",
"dependsOn": ["vpc"]
}
],
"workspace_count": 2
}
]
}
execute_workflowStarts a Terraform orchestration workflow.
Parameters:
*Either config_path or config must be provided.
Response example:
Workflow started successfully.
WorkflowID: terraform-parent-workflow-12345
RunID: abc123-def456-ghi789
get_workflow_statusGets the status of a running or completed workflow.
Parameters:
Response example:
Workflow: terraform-parent-workflow-12345
Status: WORKFLOW_EXECUTION_STATUS_COMPLETED
Started At: 2024-01-15 10:30:00
Finished At: 2024-01-15 10:35:42
The MCP server is designed for integration with AI coding assistants (like Cursor, Claude, etc.). Add it to your MCP configuration:
{
"mcpServers": {
"temporal-terraform": {
"command": "/path/to/mcp-server",
"args": []
}
}
}
infra.yaml)The configuration file defines your infrastructure workspaces and their relationships.
# Base path for resolving relative directories (optional)
workspace_root: '.'
# List of workspaces to orchestrate
workspaces:
- name: string # Required: Unique workspace identifier
kind: string # Optional: "terraform" (default and only supported value)
dir: string # Required: Path to Terraform directory
tfvars: string # Optional: Path to .tfvars file
dependsOn: [string] # Optional: List of workspace names this depends on
inputs: [InputMapping] # Optional: Variable mappings from dependencies
operations: [string] # Optional: Operations to run (default: [init, validate, plan, apply])
taskQueue: string # Optional: Override the Temporal task queue
inputs:
- sourceWorkspace: string # Name of the dependency workspace
sourceOutput: string # Name of the Terraform output to read
targetVar: string # Name of the Terraform variable to set
workspace_root: .
workspaces:
# VPC workspace - no dependencies, runs first
- name: vpc
kind: terraform
dir: terraform/examples/vpc
tfvars: terraform/examples/vpc/vpc.tfvars
dependsOn: []
taskQueue: terraform-task-queue
# Second VPC - independent, runs in parallel with first
- name: vpc-2
kind: terraform
dir: terraform/examples/vpc-2
tfvars: terraform/examples/vpc-2/vpc.tfvars
dependsOn: []
taskQueue: terraform-task-queue
# Subnets - depends on VPC, receives vpc_id output
- name: subnets
kind: terraform
dir: terraform/examples/subnets
tfvars: terraform/examples/subnets/subnets.tfvars
dependsOn:
- vpc
inputs:
- sourceWorkspace: vpc
sourceOutput: vpc_id
targetVar: vpc_id
taskQueue: terraform-task-queue
# EKS - depends on both VPC and subnets, receives outputs from both
- name: eks
kind: terraform
dir: terraform/examples/eks
tfvars: terraform/examples/eks/eks.tfvars
dependsOn:
- vpc
- subnets
inputs:
- sourceWorkspace: vpc
sourceOutput: vpc_id
targetVar: vpc_id
- sourceWorkspace: subnets
sourceOutput: subnet_ids
targetVar: subnet_ids
dependsOn)dependsOn) run in parallelinputs)The inputs array maps Terraform outputs from dependency workspaces to variables in the current workspace:
# In the subnets workspace config
inputs:
- sourceWorkspace: vpc # Get output from 'vpc' workspace
sourceOutput: vpc_id # Read the 'vpc_id' output
targetVar: vpc_id # Pass as -var vpc_id=<value>
This allows you to:
Input mappings support transitive dependencies. For example, if C depends on B, and B depends on A, then C can map outputs from both B AND A:
workspaces:
- name: a
dir: ./a
- name: b
dir: ./b
dependsOn: [a]
- name: c
dir: ./c
dependsOn: [b]
inputs:
- sourceWorkspace: a # Valid: 'a' is a transitive dependency
sourceOutput: some_output
targetVar: from_a
- sourceWorkspace: b # Valid: 'b' is a direct dependency
sourceOutput: other_output
targetVar: from_b
The operations field allows fine-grained control over which Terraform operations to run for each workspace:
operations: [init, validate, plan, apply] # Full apply mode (default)
operations: [init, validate, plan] # Plan-only mode (no apply)
Valid operations:
init - Initialize the Terraform workspace (required)validate - Validate Terraform configuration (required)plan - Generate execution planapply - Apply changes to infrastructureRequirements:
init and validate are always requiredinit → validate → plan → applyapply requires plan to be presentUse cases:
operations: [init, validate, plan] for review/approval workflowsoperations: [init, validate, plan, apply] for automatic deployments (default)workspace_root: Base path for resolving relative pathsdir and tfvars are joined with workspace_rootworkspace_root is emptyRun all tests:
go test ./...
The tests use a fake Terraform binary that:
plan (simulating changes)output and show commands.
├── activities/ # Terraform CLI wrapper activities
│ ├── terraform_activities.go # Init, Plan, Validate, Apply, Output
│ └── terraform_activities_test.go
├── cmd/
│ ├── mcp-server/ # MCP server for AI integration
│ ├── starter/ # CLI to start workflows
│ └── worker/ # Temporal worker process
├── terraform/examples/ # Sample Terraform workspaces
│ ├── vpc/
│ ├── vpc-2/
│ ├── subnets/
│ └── eks/
├── utils/ # Shared constants
├── workflow/ # Temporal workflow definitions
│ ├── config.go # Configuration types and validation
│ ├── parent_workflow.go # Orchestrator workflow
│ └── terraform_workflow.go # Per-workspace workflow
├── go.mod
├── go.sum
├── infra.yaml # Example configuration
└── README.md
Unable to create client: ...
Solution: Ensure Temporal server is running and accessible. Check TEMPORAL_ADDRESS if using a non-default location.
Invalid config: ...
Solution: Validate your infra.yaml:
dependsOn references existinputs reference valid dependenciesterraform init failed: ...
Solutions:
terraform is on PATHdir path existsplan file not found for apply: ...
Solution: This typically indicates terraform plan failed silently. Check:
Solution: The MCP server runs on stdio. Ensure your client is configured to communicate via stdin/stdout, not HTTP.
A Temporal-based workflow orchestration system for managing multi-workspace Terraform deployments with dependency resolution, variable passing between workspaces, and MCP server integration for AI-dri
Visão geral
A Temporal-based workflow orchestration system for managing multi-workspace Terraform deployments with dependency resolution, variable passing between workspaces, and MCP server integration for AI-driven automation.
Ferramentas
Ainda não há lista de ferramentas. Servidores remotos recebem uma após a primeira verificação de saúde.
Perguntas frequentes
Copie nesta página a configuração stdio para Cursor, Claude Desktop, Claude Code ou VS Code. Ela usa a instalação do publicador, não um proxy do PlugMCP.
Este registro não lista nenhuma variável de ambiente obrigatória.
Ai Driven Temporal To IAC roda localmente via stdio.
Última verificação: 2026-10-28. A Temporal-based workflow orchestration system for managing multi-workspace Terraform deployments with dependency resolution, variable passing between workspaces, and MCP server integration for AI-driven automation.