
PerfTriage
Automatically captures lag incidents on Fabric servers and points you to suspicious hotspot chunks.
Нет описания изменений
Нет описания изменений
PerfTriage 1.7.0
This update adds historical correlation to newly generated reports.
Added
- Historical correlation analysis for new lag reports
- Automatic matching against recent indexed incidents
- Repeated hotspot type counts in report output
- Repeated hotspot chunk counts in report output
- Top historical matches included in both
.txtand.jsonreports
New
-
Detection of repeated lag patterns across past incidents
-
Similar-event context directly inside newly generated reports
-
Correlation output for:
- reason
- entity hotspot type
- block entity hotspot type
- entity hotspot chunk
- block entity hotspot chunk
Changed
- Extended report generation to analyze recent indexed history before writing reports
- Added
history_correlationto the JSON report schema - Improved text reports with a dedicated
HistoryCorrelationsection
Improved
- Easier identification of recurring lag incidents
- Better visibility into repeated hotspot chunks and repeated hotspot types
- More actionable reports without manually comparing multiple files
Notes
- This update builds on the existing report index and summary system
- No new trigger logic was added; the focus is recurring-incident recognition and historical context
PerfTriage 1.6.0
This update adds report history commands and indexed summary tools.
Added
/perftriage recent/perftriage recent <limit>/perftriage summary/perftriage summary <limit>- Indexed report history reader built on top of
index.jsonl
New
- Recent incident listing directly in-game
- Historical summary of indexed lag incidents
- Breakdown of common reasons, worlds, hotspot entity types, and hotspot block entity types
- Breakdown of frequently repeated hotspot chunks
Changed
- Extended the command layer to expose report history from the existing incident index
- Improved the usefulness of indexed reports by turning them into readable operational summaries
- Made historical incident data directly accessible without opening report files manually
Improved
- Faster review of recent lag events
- Better visibility into repeated hotspot patterns
- More practical long-term troubleshooting for live servers
Notes
- This update builds on the existing report index introduced earlier
- No new lag-capture trigger logic was added in this release; the focus is historical visibility and summary workflows
PerfTriage 1.5.0
This update focuses on the report pipeline and long-term server operation.
Added
- Report index file at
config/perftriage/reports/index.jsonl - Automatic report retention cleanup
- Configurable report retention days
- Configurable maximum report file count
- Writer queue status in
/perftriage status
Changed
- Reworked the async report writer into a real bounded single-threaded queue
- Improved report write ordering so report files, index updates, and cleanup run in sequence
- Extended the config system with retention-related settings
- Updated report generation flow to append incident summaries automatically
Improved
- Better long-term behavior for servers that generate many reports
- More predictable write/load behavior under repeated lag incidents
- Reduced risk of unbounded background write buildup
Notes
- This update is focused on report infrastructure rather than new lag-detection logic
- Existing report generation behavior is preserved, while report history and cleanup are now handled automatically
- Report index file at
PerfTriage 1.4.0
A configuration-focused update that makes PerfTriage much easier to tune on live servers.
This release adds a proper config file and moves core detection / dump parameters out of hardcoded values and into user-editable settings.
New
config/perftriage/config.json/perftriage reload/perftriage config
Configurable settings
- Trigger MSPT threshold
- Trigger consecutive ticks
- Recovery MSPT threshold
- Recovery consecutive ticks
- Auto-dump cooldown
- World scan budget
- Hotspot sample count
- Hotspot Top-N size
- Hotspot per-tick budget
- Hotspot total budget
- Text / JSON report output toggles
Why this matters PerfTriage is designed for real servers, and different servers need different sensitivity and scan budgets. This update makes the mod much more practical to deploy and tune without rebuilding the jar.
Нет описания изменений
fix wrong json output
Нет описания изменений

