Skip to content
github-actions[bot] edited this page Sep 5, 2026 · 1 revision

How-To

This page collects common operator tasks using the existing repository guidance.

How to Choose a Provider

  • Choose AWS for broader enterprise patterns and scalability
  • Choose Azure for Azure-native RBAC and portal-based operations
  • Choose GCP for strong monitoring and managed service integrations
  • Choose Hetzner for the single-server hosting with operator-managed recovery

See the Provider Guide.

How to Deploy a New Instance

  1. Choose a profile from the Provider Guide
  2. Follow its README and prepare its own required variable files from that deployment directory
  3. Run terraform init
  4. Run terraform plan
  5. Run terraform apply
  6. Validate outputs and access the Foundry URL

How to Install a Game System

LegendForge does not install rulesets through Terraform.

After infrastructure is live:

  1. Log in to Foundry
  2. Install the system package that matches your Foundry version
  3. Create or restore the world
  4. Add modules gradually
  5. Test the world before inviting players

How to Operate Multiple Systems

  • Keep Foundry core upgrades controlled
  • Track which world uses which system and modules
  • Validate heavy module stacks in a test world
  • Consider separate worlds or separate deployments for very different communities
  • Back up before system or module upgrades

How to Access Running Infrastructure

AWS

Use Systems Manager Session Manager.

Azure

Use the configured administrator path or az vmss run-command invoke for the active scale set.

GCP

Use gcloud compute ssh with OS Login.

Hetzner

Use SSH directly to the server IP when enabled.

How to View Terraform State

From the provider deployment directory:

terraform state list
terraform output

Use provider-specific resource names when inspecting details with terraform state show.

How to Spin Down and Bring Back Compute

For AWS/GCP low cost, follow Low-Cost Profiles to use paused=true after first boot and backup; restore false to resume.

Persistence behavior differs by provider. Review the generated Terraform plan and the provider-specific deployment guide before disabling compute.

For Hetzner, setting:

compute_enabled = false

deletes the server and managed data volume. It is not a data-preserving pause. Use a manual server power action only after reviewing its billing and Terraform state consequences, and verify an off-server backup before any teardown-like operation. See the Hetzner lifecycle guide.

How to Prepare for Upgrades

Before major changes:

  1. Back up world data
  2. Snapshot persistent storage
  3. Record the active Foundry version
  4. Record system and module dependencies
  5. Validate in a test world first

How to Troubleshoot Safely

  • Check Terraform outputs first
  • Confirm DNS and Cloudflare Tunnel status
  • Confirm the instance or scale set is healthy
  • Check container logs
  • Check database connectivity only after application health is known

Related Pages