Satisfactory Server Performance Tuning

Satisfactory Server Performance Tuning

As your factory grows, server performance becomes critical. This guide covers the key settings to maximize stability and reduce simulation lag.

Critical: Tick Rate Lock

Satisfactory's server must run at 30 ticks per second. Higher tick rates cause instability and are not supported.

Set this in FactoryGame/Saved/Config/LinuxServer/Engine.ini:

[/Script/Engine.GameNetworkManager]
TotalNetBandwidth=104857600
MaxDynamicBandwidth=104857600
MinDynamicBandwidth=10485760

[/Game/FactoryGame/Map/GameLevel01/Persistent_Level.Persistent_Level_C]
TickInterval=0.03333

TickInterval=0.03333 = 1/30 seconds = 30 Hz.

Do not set TickInterval lower than 0.03333 (which would raise tick rate). The game engine supports 30 Hz for dedicated servers.

Network Bandwidth Settings

The TotalNetBandwidth and related settings in Engine.ini control how much data the server sends per tick. The values above (100 MB/s effective) are generally appropriate. For servers with many concurrent players:

TotalNetBandwidth=104857600      ; 100 MB/s
MaxDynamicBandwidth=104857600
MinDynamicBandwidth=10485760     ; 10 MB/s minimum

RAM Requirements by Factory Stage

The server will use more RAM as the factory grows and more machines are simulating.

Autosave Optimization

Large factories take longer to save, causing lag spikes during autosave. Tune this via the in-game Server Manager or GameUserSettings.ini:

[/Script/FactoryGame.FGGameUserSettings]
mAutosaveInterval=600    ; Save every 10 minutes instead of 5

Increasing the interval reduces save frequency but increases the risk of losing progress on crash. 5–10 minutes is a reasonable range.

Pause When Empty

Pausing factory simulation when no players are online reduces CPU and power usage significantly:

In the in-game Server Manager → Server Settings → enable Pause When Nobody Connected.

Factory Design for Server Performance

Some factory designs are more server-friendly than others:

Prefer:

• Centralized production (fewer total machine clusters)
• Manifold belt designs over complex splitter trees
• Fuel generators over biomass burners (fewer individual machines)

Avoid:

• Thousands of individually running machines (each simulates independently)
• Excessive train networks with many concurrent trains
• Large numbers of stackable conveyors going to/from lifts

Monitoring Server Load

Check CPU and memory usage:

# Real-time resource monitor
htop

# Process-specific stats
top -p $(pgrep -f FactoryServer)

# Memory usage
free -h

If CPU usage is consistently above 80%, consider:

• Pausing when empty (reduces idle load significantly)
• Reducing the number of connected players
• Moving to a higher-core-count server (factory simulation benefits from CPU speed, not just core count)