Paper Performance Settings That Matter

Paper performance tuning is strongest when you change one area at a time and measure the result. The goal is not the lowest possible settings. The goal is stable gameplay at your target player count.

#Profile Before Editing Configs

Capture evidence before changing files. If TPS drops at peak time, profile at peak time. If farms cause spikes after restarts, profile that moment. The TPS drops guide explains how to separate memory, MSPT, entities, chunks, plugins, and network symptoms.

Good tuning follows this loop:

1. Capture a profile.
2. Identify the largest tick cost.
3. Change one setting family.
4. Restart cleanly.
5. Profile the same workload again.
6. Keep or revert the change.

#Start With View and Simulation Distance

view-distance and simulation-distance shape chunk visibility and ticking. Lowering simulation distance can reduce entity and block ticking without making the world feel as visually constrained.

For many survival servers, a reasonable starting point is:

view-distance=8
simulation-distance=6

Busy servers may need lower values. Small premium servers with strong hardware can go higher if profiling supports it. Use the server.properties generator to build a baseline, then refine it per world.

#Entity Activation Ranges

Activation ranges control when entities tick at full behavior. Tightening these values helps busy survival servers, but overly aggressive settings can make farms, villagers, mobs, and redstone-dependent builds feel broken.

Focus first on entities that dominate the profile:

1. Villagers and pathfinding.
2. Item piles and hoppers.
3. Armor stands and display entities.
4. Farms in always-loaded chunks.
5. Projectiles and combat-heavy minigames.

#Hopper, Redstone, and Item Movement

Hoppers, redstone clocks, and item movement can dominate tick time in developed worlds. Do not apply broad nerfs because a public config said so. Use profiler evidence and player reports.

Safer first moves:

1. Find the region or farm creating the cost.
2. Reduce excessive always-on clocks.
3. Encourage buffered item transport.
4. Tune hopper checks carefully.
5. Avoid changes that silently break community farms without notice.

#Chunk Generation and World Borders

Chunk generation is expensive when players explore new terrain quickly. If profiles show generation cost, consider pregenerating important worlds, setting a world border, and planning exploration events away from peak hours.

Chunk send and chunk ticking are related but not identical. If players complain about missing chunks while TPS is fine, you may be looking at network, disk, or client pressure instead of server tick cost.

#Autosaves and Backups

Do not disable saves to hide lag. Data loss is worse than a short spike. Instead, tune save behavior carefully and run external backups that do not freeze the main thread.

Backups should be tested outside peak hours, stored off-node when possible, and documented so another admin can restore the server without guessing.

#JVM Flags Are Support, Not a Cure

JVM flags can reduce garbage collection problems, but they do not fix overloaded entities, heavy plugin tasks, or extreme view distance. If your evidence points to garbage collection, generate a conservative command with the Minecraft JVM flag generator. If the evidence points to tick work, tune the workload first.