Unless something actually crashes, which any code could do, then why would you need to reload? Just send a websocket message that requests the current full state.
But if it's implemented correctly, it probably won't get out of sync.
Remote code doesn’t load correctly sometimes, just a fact of life in a distributed system. Maybe loading it locally reduces that to almost zero, but I’d still be more comfortable risking my life with a desktop app.
But if it's implemented correctly, it probably won't get out of sync.