Performance optimization guide

Right-size your plan

• Track average CPU, RAM and disk usage during peak hours. Upgrade before you sustain 80% utilization.
• For memory-hungry workloads, choose plans with higher RAM-to-core ratios and avoid swapping.
• Move data-heavy servers to NVMe plans for better I/O throughput.

Unsure which plan fits? Open a ticket with recent metrics and the reliability team will recommend a configuration.

Tune the JVM and runtime flags

Many game servers and mod loaders run on the JVM. Adjust launch flags from the schedules or start commands.

• Use Aikar's flags for Minecraft Paper/Purpur instances to balance CPU usage and GC pauses.
• Set -Xms and -Xmx to the same value to avoid heap resizing overhead.
• Enable -XX:+ZGC or -XX:+UseG1GC for large heaps (>8 GB).
• Log garbage collection with -Xlog:gc* and review output weekly.

Reduce tick and TPS pressure

• Cap mob spawning and redstone updates on survival servers.
• Pre-generate chunks and store them on NVMe-backed storage to reduce on-demand generation.
• Disable unused plugins or modules—each one consumes CPU even when idle.
• Profile with spark or WarmRoast and fix the top offenders before the next event.

Network optimization

VaultScope terminates DDoS mitigation in front of your server. You still control ports and subdomains via the panel.

• Keep only required ports open. Reclaim extras to avoid scanning noise.
• Use subdomains from vaultscope.dev or bring a reverse proxy to manage SSL certificates.
• Place latency-sensitive servers in the region closest to your players. Use health checks to detect when a region degrades.

Caching and external services

• Offload static assets (resource packs, maps) to a CDN or object storage.
• Use Redis or Memcached for frequently accessed data rather than hitting the database every tick.
• Enable connection pooling and configure sensible timeouts to prevent resource leaks.

For complex setups, consider containerized sidecars or separate compute nodes connected via the VaultScope network.

Load testing and validation

1. Clone production to a staging server using backups.
2. Replay traffic with tools like tcpreplay, JMeter or custom bots.
3. Monitor metrics in real-time and annotate when you apply changes.
4. Document results and adjust runbooks before rolling updates to production.