Skip to main content
The Spark Profile Analyzer automatically analyzes your server’s performance profile and provides specific optimization recommendations. Access it at birdflop.com/resources/sparkprofile.

​
What is Spark?

Spark is a performance profiler plugin for Minecraft servers that helps identify:
  • CPU usage by plugins and game systems
  • Tick duration and performance bottlenecks
  • Thread activity and resource consumption
  • Memory allocation patterns

​
Getting Started

1

Install Spark

Download and install the Spark plugin from spark.lucko.me on your server.
Purpur servers 1.19+ have Spark built-in, no installation needed.
2

Start Profiling

Run the command in-game or console:
/spark profiler
This begins recording server performance data.
3

Stop and Get Link

After 30-60 seconds of gameplay, stop the profiler:
/spark profiler --stop
Copy the link provided (looks like https://spark.lucko.me/xxxxxx).
4

Analyze

Paste the Spark link into the analyzer at birdflop.com/resources/sparkprofile and click Submit.

​
Understanding the Analysis

The analyzer examines your profile and provides recommendations in several categories:

​
Plugin Performance Issues

Identifies plugins consuming excessive CPU time:
  • Plugins using >10% of tick time
  • Inefficient event handlers
  • Poorly optimized operations
  • Suggestions for alternatives or configuration changes

​
Server Configuration

Recommends configuration changes:
  • View distance adjustments
  • Entity limits and spawning
  • Chunk loading settings
  • Redstone optimization
  • Mob AI settings

​
World Issues

Detects world-specific problems:
  • Excessive entities in specific chunks
  • Overloaded tile entities (hoppers, furnaces, etc.)
  • Chunk generation lag
  • Redstone contraptions causing lag

​
System Performance

Analyzes system-level issues:
  • CPU bottlenecks
  • Memory pressure
  • Disk I/O problems
  • Java version recommendations

​
Reading Recommendations

Each recommendation includes:
  1. Severity Level: Critical, High, Medium, or Low priority
  2. Description: What the issue is and why it matters
  3. Solution: Specific steps to resolve the issue
  4. Impact: Expected performance improvement
These are not magic values. Many settings have real consequences on your server’s mechanics. See this guide for detailed information.

​
Best Practices

Profile During Peak Activity: Run the profiler when your server is experiencing lag for the most useful data.
Profile for 30-60 Seconds: Longer profiles provide more accurate data, but 30-60 seconds is usually sufficient.
Profile Different Scenarios: Create separate profiles for different situations (player events, chunk generation, etc.).

​
When to Profile

  • During player-reported lag
  • After installing new plugins
  • When testing configuration changes
  • During regular maintenance checks
  • Before and after optimization attempts

​
What to Look For

In the Spark web viewer, pay attention to:
  1. Top Contributors: Plugins/systems at the top of the flame graph
  2. Tick Duration: Should be under 50ms (20 TPS)
  3. GC Activity: Frequent garbage collection indicates memory issues
  4. Thread Count: Too many threads can cause problems

​
Common Issues and Solutions

​
High Tick Time

Symptom: Server TPS below 20, laggy gameplay Common Causes:
  • Too many entities in loaded chunks
  • Inefficient plugins (especially anticheats)
  • Large redstone contraptions
  • Excessive chunk loading
Solutions:
  • Reduce entity spawning in server.properties and paper-world-defaults.yml
  • Optimize or remove problematic plugins
  • Limit redstone clock speeds
  • Reduce player view distance

​
Plugin Overhead

Symptom: Specific plugin consuming >10% of tick time Solutions:
  • Check plugin configuration for optimization options
  • Update to latest version (may include performance fixes)
  • Look for alternative plugins
  • Contact plugin developer with profiler data

​
World Issues

Symptom: Lag in specific areas or chunks Solutions:
  • Use /spark tickmonitoring to find laggy chunks
  • Reduce entities in problem areas
  • Remove excessive tile entities (hoppers, etc.)
  • Consider regenerating extremely problematic chunks

​
Memory Issues

Symptom: Frequent garbage collection in profiler Solutions:
  • Allocate more RAM to your server
  • Optimize JVM flags (see Flags Generator)
  • Reduce loaded chunks (view distance, simulation distance)
  • Remove memory-heavy plugins
You can skip the input form by constructing a direct link: 
https://birdflop.com/resources/sparkprofile/[code]
Replace [code] with the Spark profile ID from the spark.lucko.me URL.

​
Advanced Profiling

​
Thread Profiling

Profile specific threads: 
/spark profiler --thread [thread-name]

​
Sampling Interval

Adjust sampling rate (lower = more detailed, higher overhead): 
/spark profiler --interval 4

​
Focus on Specific Plugin

Profile only a specific plugin: 
/spark profiler --only-ticks-over 50

​
Limitations

  • Profiling adds ~5% overhead while active
  • Short profiles may not capture intermittent issues
  • Cannot profile startup lag (only runtime)
  • Some native code operations may not appear in profile

​
Powered By Botflop

This analyzer uses the same analysis engine as the Botflop Discord bot, which can automatically analyze profiles posted in Discord.
Add Botflop to your Discord server to get instant profile analysis when you share Spark links in chat.

Build docs developers (and LLMs) love

Get started for freeTalk to us