Direct answer

A mob farm has multiple bottlenecks: spawn attempts, live mob AI, killing, XP/items and hoppers. Identify which stage is slow rather than lowering the entire server mob cap.

How to check it

  1. On Paper use /paper mobcaps and /paper playermobcaps <player> to inspect natural mob populations. A full cap explains poor natural spawning but not necessarily high MSPT.
  2. Run /spark profiler start --timeout 300 while the farm is running. Note whether entity ticking, block updates, item entities or plugin hooks dominate.
  3. Test idle, spawning, killing and collection phases separately. Record AFK position, dimension and whether other players are competing for spawn opportunities.

What to change

  1. Prevent mobs and items from queuing indefinitely; a kill chamber filled with living mobs often costs more than a smooth flow of quickly processed spawns.
  2. Reduce unnecessary hopper clocks/transport where profiling shows block updates are expensive. Use changes that preserve collection reliability.
  3. Changing global spawn limits affects every farm and natural world activity. Prefer correcting the farm bottleneck first, then consider documented world-specific configuration.

Software and version notes

Version and mechanics warning

Natural spawning and spawner-cage spawning follow different rules. Do not apply a natural mob-cap diagnosis to a spawner farm without checking its mechanism.

Verify the result

Track loot throughput, active mobs and MSPT for the same number of players. Confirm normal natural spawning elsewhere remains acceptable.

References