Confirm what is actually slow
Server TPS measures how quickly the server advances the game simulation; 20 TPS is the normal target. Client FPS is rendered on the player’s computer, while network latency is a separate connection problem.
A “Can’t keep up” warning means the server missed its tick deadline. It is useful evidence of server-side delay, but it does not by itself prove that the CPU is the only bottleneck. Heavy chunk generation, garbage collection, disk stalls, plugins, mods and other work can also make ticks late.
Measure before buying more RAM
- Check recent console output and resource graphs around the slowdown.
- Use the validated Paper/Purpur profiling workflow when available to find expensive plugins, entities, chunks or tasks.
- Separate sustained near-limit memory or an OutOfMemoryError from normal heap use and garbage collection.
- Check whether the problem appears during exploration, backups, world saves or a specific scheduled/plugin task.
- Change one major variable at a time so you can tell what actually helped.
ImportantMore RAM helps when the server is genuinely memory-constrained. It does not automatically make a tick-heavy or I/O-bound server faster.
Reduce common tick costs
- Lower view-distance and simulation-distance to reasonable values for the community.
- Control excessive entities, item piles, mob farms, redstone, hoppers and automation that stay active every tick.
- Remove, update or reconfigure plugins and mods that profiling identifies as expensive.
- Pre-generate the intended world area when exploration-time chunk generation is a major source of stalls.
- Keep the server software and supported plugins/mods on compatible versions rather than stacking unrelated optimization tweaks.
Protect the world while tuning
Create a recovery point before large plugin, mod, configuration or world-generation changes. If an optimization causes startup errors, broken mechanics or world issues, roll back the last change instead of layering several more changes on top of it.