Issue with fly proxy not routing to last running machine in multi-region phoenix app

I have a Phoenix app running on up to 6 machines across 3 regions, with the main region being AMS. The configuration ensures that at least one machine is always running (min_machines_running = 1). Currently, I’m facing an issue where fly proxy, instead of routing to the last running machine, chooses to start a stopped one. This occurs with both standard HTTP requests and WebSocket connections, used by Phoenix.LiveView. The restart of the stopped machine takes around 3 seconds or more. Any suggestions on what might be causing this behavior? Is there a way to configure fly proxy to route to the last running machine? Or do I need to keep one machine running in each region?

This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.