How to Install and Use Spark for Minecraft Performance Checks

Running a heavily modded Minecraft server can be great until TPS starts sinking and everyone begins blaming the nearest chicken farm. Lag can come from entities, chunks, redstone, plugins, mods, memory pressure, CPU load, or several of those at once. Spark helps narrow that down.

Spark is a performance profiling mod for Forge and Fabric. It can run reports for a dedicated server or a singleplayer client, showing information such as TPS, tick duration, CPU usage, memory usage, and network details. The reports are especially useful when a server feels slow but the cause is not obvious.

Download Spark

• Open the Spark page on CurseForge and select Files near the top.

• Find the Spark version that matches your Minecraft version and mod loader.
• Click the three-dot menu beside the file, then choose Download File.

• Save the `.jar` somewhere easy to find, such as your downloads folder or desktop.

Install Spark on Your Client

Spark needs Forge or Fabric installed in the Minecraft Launcher before it can load. Client installation is optional if you only want to diagnose a dedicated server, but it is useful for singleplayer worlds or local testing.

• Open the Minecraft Launcher and go to Installations.

• Locate your Forge or Fabric profile, then click the folder icon.

• Open the `mods` folder. If it does not exist, create one.
• Move the Spark `.jar` file into that folder, then return to the launcher.

• Start the modded profile with Play to load Spark.

Install Spark on a Server

For server-side troubleshooting, your server must already be running Forge or Fabric. Make sure the loader matches the Spark file you downloaded, then restart once if needed so the server creates its required folders.

• Open your server control panel and go to the file manager or connect with FTP.

• Sign in if required, then open the `mods` directory.

• Upload the Spark `.jar` file into the `mods` folder.

• Wait for the upload to finish, then restart the server.

Using Spark Commands

Spark is controlled through commands, so you need cheats in singleplayer or operator permissions on a server. Once you have permission, run `/spark help` to view the available commands.

The mod has many tools, but the profiler and health report are the two most useful starting points. They give enough detail to confirm whether the server is actually struggling or just having one dramatic moment.

Run the Profiler

The Spark profiler records what is using server resources and produces a detailed report. This can include tick data, entity activity, CPU usage, and memory usage with values and percentages.

Use this command to start or stop it:

`/spark profiler start`

`/spark profiler stop`

Let the profiler run for about 5 to 15 minutes before stopping it, especially if the lag happens during normal gameplay. If you want it to stop automatically, add a timeout in seconds:

`/spark profiler start --timeout [seconds]`

After the profiler stops, Spark provides a report link you can review or share with someone helping diagnose the issue.

Check Server Health

For a faster overview, use Spark's health report command:

`/spark healthreport --network`

Some versions may use this instead:

`/spark health`

This shows current TPS, tick duration, CPU load, memory usage, and network information. It is a good first check before running a longer profiler session, since it can quickly show whether the server is under pressure.

Other Useful Spark Tools

Spark includes smaller commands such as `/spark ping`, `/spark tps`, and `/spark tickmonitor`. These are not as detailed as the profiler, but they are handy for quick checks while players are online.

For advanced usage, the official Spark documentation is worth keeping nearby. The command list is more interesting than it sounds, which is a low bar for documentation but still useful.

Why Spark Helps Optimization

Spark does not magically fix lag by itself. Instead, it shows where the problem is likely coming from. A report might point toward oversized mob farms, expensive mods, overloaded chunks, memory pressure, or a server that simply needs more resources.

With that information, you can make focused changes instead of guessing. Remove or tune the problem mod, adjust farms, reduce entity buildup, review startup flags, or upgrade resources when the data supports it.

Useful Links

• Spark on CurseForge
• Official Spark documentation
• How to install Minecraft mods
• Modded Minecraft server optimization