Connectivity Problem with Minecraft - Minecraft server experiencing persistent severe TPS lag

Technical Support
, ,
1

System Information

Field Value
Operating System Linux - Debian GNU/Linux 12 on x86_64
Product AMP ‘Phobos’ v2.6.3.2 (Mainline)
Virtualization Docker
Application Minecraft
Module MinecraftModule
Running in Container Yes
Current State Ready

Problem Description

Issue

The Minecraft server is suffering from chronic and severe TPS (Ticks Per Second) lag. This issue is persistent and makes the server nearly unplayable for users.

Symptoms include:

The server console is continuously flooded with warnings like: [Server thread/WARN]: Can’t keep up! Is the server overloaded? Running 5000ms or 100 ticks behind.

The actual TPS remains very low, usually between 5 and 15.

This lag occurs even with a minimal player count (e.g., 1 or 2 players) and sometimes even when the server is empty.

Troubleshooting Performed:

I have checked the resource usage in the Hostinger control panel, and the CPU and RAM usage do not appear to be maxed out when the lag occurs. This suggests the bottleneck might be related to single-core performance, disk I/O speed, or limits on resource allocation imposed on the container.

I require assistance from your side to investigate the server environment and determine the root cause of the high tick usage.

Reproduction Steps

  • Start the Minecraft server (using PaperMC/Spigot on version 1.21.1).
  • Observe the server console, where warnings regarding “Running XXX ticks behind” appear consistently after the server finishes starting.
  • Check the server’s TPS (Ticks Per Second) using the /tps command, which consistently reports well below the optimal 20.0 (e.g., 5-15 TPS).
2

Please note that this is the forum for AMP, and not Hostinger support, so we don’t have access to the hardware you’re using. We’ll still try and get you pointed in the right direction though.

You can use a plugin like Spark to help you narrow down TPS drops. Commonly people set the server-side chunkloading too high, and since chunkloading is an expensive (mostly-singlethreaded) operation, the server’s TPS tanks. The default for this setting is 10.

Sometimes disabling the Use Synchronous IO can help, but it’s better to take a read through the Spark profiler to see what the root cause is.