Server Performance Optimization

Hytale Server Performance Optimization

Optimizing your Hytale server performance is crucial for providing a smooth, lag-free experience for your players. This comprehensive guide covers all aspects of server optimization from basic settings to advanced tuning.

Performance Factors Overview

Server performance depends on these key factors:

• CPU processing power and core allocation
• RAM usage and garbage collection efficiency
• Disk I/O speed and storage type
• Network bandwidth and latency
• World size and player activity

Solution 1: Hardware Optimization

CPU Configuration

Optimize CPU usage for server workloads:

# Java CPU affinity and threading
-XX:+UseNUMA
-XX:+UseParallelGC
-XX:ParallelGCThreads=4
-XX:ConcGCThreads=2
-XX:+UnlockExperimentalVMOptions
-XX:+UseTransparentHugePages

Resource Allocation Guidelines

Solution 2: Java Virtual Machine Tuning

Garbage Collector Selection

Choose the right GC for your server size:

# Small servers (1-20 players)
-XX:+UseSerialGC
-XX:MaxGCPauseMillis=200

# Medium servers (20-100 players)
-XX:+UseG1GC
-XX:MaxGCPauseMillis=150
-XX:G1HeapRegionSize=16M

# Large servers (100+ players)
-XX:+UseZGC
-XX:ZCollectionInterval=30

Memory Optimization

Advanced memory tuning parameters:

-Xms6G -Xmx12G
-XX:+UseCompressedOops
-XX:+UseCompressedClassPointers
-XX:+UseStringDeduplication
-XX:+OptimizeStringConcat
-XX:NewRatio=2
-XX:SurvivorRatio=8

Solution 3: Server Configuration

World Generation Settings

The setting with the most effect is MaxViewRadius in config.json: the maximum view radius in chunks (a chunk is 32 blocks wide), 32 by default and accepted from 1 to 32. A smaller radius means fewer chunks loaded and sent per player.

{
    "MaxViewRadius": 12
}

Each world's own config.json also has IsSpawningNPC, IsBlockSpawnersEnabled and IsTicking, and /world settings changes them live. /world perf shows tick performance and /world tps sets the tick rate for a world.

Network Configuration

Hytale uses QUIC over one UDP port, so there are no TCP, buffer or compression settings to tune in the server file. The network setting that exists is the packet rate limit in config.json:

{
    "RateLimit": {
        "Enabled": true,
        "PacketsPerSecond": 2000,
        "BurstCapacity": 500
    }
}

Those are the built-in defaults: 2,000 packets per second per client with a burst of 500. Raise them only if players on busy builds are being dropped, and leave them alone when nothing is wrong.

Solution 4: Plugin and Mod Optimization

Performance Monitoring

Track plugin performance impact:

• Use profiling tools to identify resource-heavy plugins
• Monitor plugin execution times
• Check for memory leaks in custom code
• Benchmark plugin performance with and without each mod

Plugin Optimization Best Practices

Solution 5: System-Level Optimization

Operating System Tuning

Hytale talks to players over QUIC on one UDP port, so the TCP buffer and congestion settings in generic Linux tuning guides do nothing for its game traffic. The setting that matters is the UDP receive buffer. The server (read from the server files of build 0.7.3) asks the kernel for a 33,554,432-byte (32 MiB) receive buffer on its IPv4 and IPv6 sockets, and when it gets less than 25,165,824 bytes (24 MiB), it logs this at every start:

UDP receive buffer is 425984 bytes (effective 212992), below the 25165824-byte autotune window. Raise net.core.rmem_max so inbound bursts are not dropped

Most Linux systems cap the buffer at 212,992 bytes. Raise the cap to what the server asks for, then restart the server, because it sizes the buffer when it opens its port:

# /etc/sysctl.d/99-hytale.conf, applied with: sudo sysctl --system
net.core.rmem_max = 33554432

Process Priority Management

Set appropriate priorities for server processes:

# Set high priority for server process
nice -n -10 java [server_arguments]

# CPU affinity for dedicated cores
taskset -c 0-3 java [server_arguments]

# I/O priority for better disk performance
ionice -c 1 -n 7 java [server_arguments]

Performance Monitoring

Key Metrics to Track

Monitoring Tools

Use these tools for comprehensive monitoring:

• Prometheus + Grafana: Advanced metrics visualization
• Java Mission Control: JVM performance analysis
• htop/nmon: System resource monitoring
• Wireshark: Network traffic analysis
• Custom Scripts: Automated performance alerts

Troubleshooting Performance Issues

Common Performance Problems

Performance Testing

Regular performance testing methodology:

1. Establish baseline performance metrics
2. Simulate typical player load patterns
3. Test with stress scenarios (maximum capacity)
4. Monitor resource usage during testing
5. Document results and compare over time

Advanced Optimization Techniques

Load Balancing

For very large servers, consider these strategies:

• Distribute world processing across multiple instances
• Use proxy servers for connection handling
• Implement database clustering for data storage
• Separate chat and auxiliary services

Caching Strategies

The server has no cache layer to configure. Its storage is set in config.json through PlayerStorage, UniverseResourceStorage and BanStorage providers, and world chunks use the storage type in each world's ChunkStorage setting. Change these only on advice from the server documentation.

Maintenance and Updates

Regular Optimization Tasks

• Weekly performance review and metric analysis
• Monthly JVM tuning adjustments
• Quarterly hardware assessment and upgrades
• Continuous plugin optimization and updates
• Regular system maintenance and cleanup

Maximize your Hytale server performance with proper optimization, monitoring, and maintenance to provide the best possible experience for your community.