Performance
Servers → Performance (/p/<projectId>/servers/performance)
Why this page exists
Performance helps you answer: Is the server struggling before players complain? Use it to connect lag reports to TPS, CPU, and memory, and to see whether world size (chunks, entities) lines up with load. Good for capacity planning, finding the noisy world, and proving an upgrade helped.
How to use it
Select one server from the dropdown (each server has its own graph; there is no combined view). The list defaults to the first server when it loads.
Pick a date range that matches the incident (for example last 24 hours for spikes, or several days to see a pattern). Your plan may cap history.
Scan the summary cards for averages: TPS, CPU, memory, and peak players in the range. Use these as headline health before the charts.
Read TPS over time first: sustained low TPS usually means tick lag for players.
Compare CPU and memory charts to see if the host is compute-bound or RAM-bound.
Check world stats (chunks, entities, players per world when shown) when TPS is bad but CPU looks fine: heavy worlds or entity farms are common causes.
What to do with the insights
TPS drops at peak hours: you may need more hardware, fewer entities, or lighter plugins; see Plugins for version drift.
CPU and memory high together: overall overload; profile plugins and world size.
Memory high but CPU moderate: look at allocations, chunk loading, or leaks (plugin updates).
If you see no data
Performance needs status snapshots from PlayerAtlas on that server. Confirm the plugin is installed, online, and reporting. If the server was off for the whole range, the page will stay empty.
Notes
Charts bucket data into 5-minute windows when multiple samples arrive close together. TPS is a Minecraft-oriented metric; other games may report different signals in the same area as the product evolves.
Last updated