Configuration

spark has a limited set of configuration options. They are quite niche and most users will never need to change from the defaults!

To override the default configuration values, you can either:

• create a configuration file, config.json, in the spark plugin/mod folder (the config file is not automatically generated)
• set environment variables
• set a Java system properties

Options

Below is a list of the available configuration settings.

---

backgroundProfiler

If background profiling is enabled.

When true, spark will automatically start profiling in the background when the server/client/proxy starts up. The profile can be uploaded using the usual commands. The default value is true.

Example (config file)

{
    "backgroundProfiler": true
}

Example (environment variable)

export SPARK_BACKGROUNDPROFILER=true

Example (system property)

java -Dspark.backgroundProfiler=true

---

backgroundProfilerInterval

The interval that the background profiler should sample at. The default value is 10 (milliseconds).

Example (config file)

{
    "backgroundProfilerInterval": 10
}

Example (environment variable)

export SPARK_BACKGROUNDPROFILERINTERVAL=10

Example (system property)

java -Dspark.backgroundProfilerInterval=10

---

backgroundProfilerEngine

The engine that the background profiler should use.

The default value is "async", but "java" is also supported.

Example (config file)

{
    "backgroundProfilerEngine": "async"
}

Example (environment variable)

export SPARK_BACKGROUNDPROFILERENGINE=async

Example (system property)

java -Dspark.backgroundProfilerEngine=async

---

viewerUrl

The URL used when providing a link to the viewer in command outputs.

The data "code" is appended after the URL, so it should always end with a /.

The default value is "https://spark.lucko.me/".

Example (config file)

{
    "viewerUrl": "https://spark.lucko.me/"
}

Example (environment variable)

export SPARK_VIEWERURL=https://spark.lucko.me/

Example (system property)

java -Dspark.viewerUrl=https://spark.lucko.me/

---

bytebinUrl

The URL for the bytebin instance which profiles & heap dump summaries should be uploaded to. The URL should end with a /.

The default value is "https://spark-usercontent.lucko.me/".

Example (config file)

{
    "bytebinUrl": "https://spark-usercontent.lucko.me/"
}

Example (environment variable)

export SPARK_BYTEBINURL=https://spark-usercontent.lucko.me/

Example (system property)

java -Dspark.bytebinUrl=https://spark-usercontent.lucko.me/

---

bytesocksHost

The hostname for the bytesocks instance which should be used for communicating with the spark viewer.

The default value is "spark-usersockets.lucko.me".

Example (config file)

{
    "bytesocksHost": "spark-usersockets.lucko.me"
}

Example (environment variable)

export SPARK_BYTESOCKSHOST=spark-usersockets.lucko.me

Example (system property)

java -Dspark.bytesocksHost=spark-usersockets.lucko.me

---

overrideTpsCommand

If spark should override the default TPS command with its own. This setting is only read when spark is running on a Bukkit server.

The default value is true.

Example (config file)

{
    "overrideTpsCommand": true
}

Example (environment variable)

export SPARK_OVERRIDETPSCOMMAND=true

Example (system property)

java -Dspark.overrideTpsCommand=true

---

disableResponseBroadcast

If spark shouldn't broadcast command output to all online administrators (players with the necessary permissions to use spark).

The default value is false.

Example (config file)

{
    "disableResponseBroadcast": false
}

Example (environment variable)

export SPARK_DISABLERESPONSEBROADCAST=false

Example (system property)

java -Dspark.disableResponseBroadcast=false

---

maxStackDepth

The max depth of stack frames recorded by the sampler. Any frames over this depth are discarded/ignored.

Setting to a modest figure ensures that excess data is not needlessly recorded and prevents stack overflow errors if downstream processes are reading the data via recursive method calls.

The default value is 300.

note

This setting can only be configured using a system property.

Example (system property)

java -Dspark.maxStackDepth=300

---

continuousProfilingHistorySize

The number of windows to record in continuous profiling before data is dropped. Each window is approximately 1 minute.

The default value is 60.

The official spark viewer does not accept uploads above 10MB in size, and setting this value above 60 is more likely to breach this limit. If you run into HTTP 413 errors, try reducing this value.

note

This setting can only be configured using a system property.

Example (system property)

java -Dspark.continuousProfilingHistorySize=60

---

useOsTmpDir

Whether to use the built-in operating system's temporary directory to store temporary spark files.

The default value is false.

useOsTmpDir setting	OS	Likely path
true	Windows	C:\Users\<username>\AppData\Local\Temp\
true	macOS	/var/folders/.../T/
true	Linux	/tmp/
false	n/a	plugins/spark/tmp (or similar)

note

This setting can only be configured using a system property.

Example (system property)

java -Dspark.useOsTmpDir=true

---