Windrose Server High CPU Usage When Empty | Known Bug & Workarounds

If your Windrose dedicated server experiences abnormally high CPU usage when no players are connected, you are encountering a known Early Access bug. This issue causes the server to saturate two CPU cores when empty, but CPU usage drops significantly once a player joins. We're also seeing this on servers hosted here at HypeServ!

Understanding the Bug

Unlike standard Unreal Engine 5 servers that enter a sleep state, the Windrose server currently fails to hibernate correctly when the world is empty. Because the simulation loops and NPC AI continue running at full speed, the server uses more processing power when idle than when active. This inverted behavior is present across all hosting environments, including HypeServ, Docker, and direct self-hosting.

Hardware Impact

This bug has significant hardware implications because it fully saturates the minimum official requirement of two CPU cores whenever the server is empty. Since the server does not scale down its processing, instances running your local PC will use significant performance and certain cloud environments may trigger an alarm.

Practical Workarounds

Schedule automatic server shutdowns during known offline hours using Windows Task Scheduler, Linux cron jobs, or scheduled tasks in the HypeServ control panel.
Assign lower CPU weights in Docker Compose to deprioritize the idle Windrose server when other active containers need processing power.
Update your server binary regularly to ensure you receive the official hibernation patch from Kraken Express when it releases.

Expected Future Patches

Kraken Express is expected to release an update that implements proper server hibernation and tick rate throttling. They have already released hotfixes addressing dedicated server connectivity and version mismatch issues, indicating active development on server-side performance. Once the hibernation fix is applied, idle CPU consumption should drop below five percent. Until this optimization arrives, applying resource caps or scheduling downtime remains the most effective strategy.