Settings enabling controllers and hosts to function efficiently

[alameth]

OS Name/Version:
Debian 12 (Bookworm) – fully updated
Version : 12.13

Product Name/Version:
AMP 2.6.0.10 (latest stable build)

Problem Description:
I was previously running a standalone AMP instance on a single Linux VM to manage 4–5 game servers (Minecraft, Hytale, etc.) without any issues.

After migrating my infrastructure for better resilience and to avoid downtime during maintenance, I switched to a distributed AMP setup consisting of 3 VMs:

• 1 Controller
• 2 Target hosts

All 3 VMs are on the same network, built using the same deployment script and identical configuration parameters.

However, when managing instances on the targets from the Controller, action buttons (start/stop/restart) do not respond, and the integrated game consoles (for Minecraft, Hytale, etc.) are not functional.

Both connection modes — “via Target” and “Direct From Controller” — were tested, with the same results.

It seems I may have missed or misconfigured something in the controller/target communication, but I cannot identify what.

I’d appreciate your expertise on the proper configuration and communication flow between Controller and Targets, or hints on what could block management/console functionality in this setup.

Steps to reproduce:
Step 1: Deploy three Debian 12 VMs (one controller, two targets) using the same deployment script and identical AMP setup configuration.
Step 2: Register both targets into the controller using AMP’s UI.
Step 3: Attempt to control or view a game instance (e.g., Minecraft or Hytale) from the controller’s dashboard.
Step 4: Observe that action buttons and consoles are unresponsive or not loading properly.

Actions taken to resolve so far:

• Verified system clocks, DNS resolution, and hostname reachability between all VMs (OK)
• Confirmed controller–target connectivity (ping OK, SSH OK)
• Network verification:
  • Verified AMP ports open (8080, 8081, 8082, 8083, 2223) on all VMs: netstat -tlnp | grep amp
  • Checked firewall rules: ufw status and iptables -L — all AMP ports allowed
  • Tested port connectivity between controller and targets: nc -zv target1 8080/8081/8082/8083
  • Rebound AMP services: systemctl restart amp on all nodes
• Browser console check: F12 Developer Tools → Console/Network tabs show no errors, but WebSocket connections to targets fail/time out
• Tested both connection modes (“Via Target” and “Direct From Controller”) without success
• Restarted AMP services and instances on both controller and targets
• Checked AMP logs (~/AMP/Logs) on all nodes for authentication/communication errors (no critical errors)
• Reinstalled AMP on a fresh VM using the same procedure to rule out corruption — same behavior
• Test EnablePassthruAuth=true
• Check controller’s URL