Diagnostics

/rtp info is the operator diagnostics dashboard. It reports loaded worlds, permanent regions, queue statistics, and learned-state (spatial memory) summaries - the live view of whether the engine is keeping up. The charts and counters render from in-memory state, so reading them never touches the teleport hot path.

Happy path

/rtp info                  # whole-server overview (permission: rtp.info)
/rtp info region=<name>    # focus a single region
/rtp info world=<name>     # focus a single world

What the dashboard shows

Group	What it tells you
Worlds & regions	Which worlds are loaded and which permanent regions exist.
Cache depth	Ready-to-serve locations vs capacity (cached / keptCache / unkeptCache / cacheCap). A healthy region keeps this near its cap.
Backlog	Unverified backlog buffer depth (backlogCache / backlogCacheCap).
Queue & in-flight	Players waiting on a coordinate (locationQueue) and calculations in progress (inFlightCalculations).
Pipeline latency	Teleport-pipeline percentiles (pipelineMsP50 ... pipelineMsP99) and the slow-pipeline count over the configured threshold.
Spatial memory	Coverage and bad-location stats (memCoveragePct, memBadPct, memBadCount) and the top rejection cause (memTopCause / memTopCausePct).
TPS / MSPT	Server tick health; on Folia, a per-region breakdown.

Reading it

• Cache sitting near zero with a non-zero queue means generation is not keeping up with demand. Run /rtp scan to build spatial memory, and review performance.yml (attempt limits, cache rate, cacheCap).
• High memBadPct / a dominant memTopCause tells you why candidates are being rejected (biome, unsafe block, claim). If a biome or block filter is too aggressive, adjust safety.yml or the region's biome controls.
• Rising pipeline percentiles point at chunk-load or verification cost; confirm the Anvil pre-filter is doing its job and that spatial memory is populated.
• Low TPS / high MSPT during /rtp bursts is the symptom the engine is designed to avoid; if you see it, capture /rtp info output and a timings report for a bug report.

Exposing the same data elsewhere

Every counter above is also available as a PlaceholderAPI placeholder (%rtp_<key>%) and inside messages.yml as [key], so you can surface queue depth or latency on a scoreboard or admin panel without /rtp info. The metrics snapshot knobs (TPS/MSPT/heap aggregation, mostly Folia) live in metrics.yml.

See also: Performance, Metrics, PlaceholderAPI, Scan and Spatial Memory.