WildFly JSON Formatter

Following settings can be used:

Attribute Name	Description	Default
lineBreak	End of line string	`\n`
fields	Comma-separated list of log event fields that should be included in the JSON	`Time, Severity, ThreadName, SourceClassName, SourceMethodName, SourceSimpleClassName, LoggerName, NDC, Server`
version	GELF Version `1.0` or `1.1`	`1.0`
originHost	Originating Hostname	FQDN Hostname
extractStackTrace	Send the Stack-Trace to the StackTrace field (`true`/`false`)	`false`
filterStackTrace	Perform Stack-Trace filtering (`true`/`false`)	`false`
includeLogMessageParameters	Include message parameters from the log event	`true`
includeLocation	Include source code location	`true`
facility	Name of the Facility	`logstash-gelf`
mdcProfiling	Perform Profiling (Call-Duration) based on MDC Data. See MDC Profiling for details	`false`
additionalFields	Send additional static fields. The fields are specified as key-value pairs are comma-separated. Example: `additionalFields=fieldName=Value,fieldName2=Value2`	none
additionalFieldTypes	Type specification for additional and MDC fields. Supported types: `String`, `long`, `Long`, `double`, `Double` and `discover` (default if not specified, discover field type on parseability). Eg. field=String,field2=double	`discover` for all additional fields
mdcFields	Send additional fields whose values are obtained from MDC. Name of the Fields are comma-separated. Example: `mdcFields=Application,Version,SomeOtherFieldName`	none
dynamicMdcFields	Dynamic MDC Fields allows you to extract MDC values based on one or more regular expressions. Multiple regexes are comma-separated. The name of the MDC entry is used as GELF field name.	none
dynamicMdcFieldTypes	Pattern-based type specification for additional and MDC fields. Key-value pairs are comma-separated. Example: `my_field.=String,business\..\.field=double`	none
includeFullMdc	Include all fields from the MDC.	`false`
timestampPattern	Date/time pattern for the `Time` field	`yyyy-MM-dd HH:mm:ss,SSS`

The JSON formatter creates then messages like:

{
    "LoggerName": "org.wildfly.extension.undertow",
    "short_message": "WFLYUT0008: Undertow HTTP listener default suspending",
    "level": "6",
    "Time": "2015-08-11 21:25:32,0734",
    "Severity": "INFO",
    "MessageParam1": "default",
    "Thread": "MSC service thread 1-6",
    "SourceMethodName": "stopListening",
    "MessageParam0": "HTTP",
    "full_message": "WFLYUT0008: Undertow HTTP listener default suspending",
    "SourceSimpleClassName": "HttpListenerService",
    "SourceClassName": "org.wildfly.extension.undertow.HttpListenerService",
    "facility": "logstash-gelf",
    "timestamp": "1439321132.734"
}

Default line break is \n, JSON objects are separated by line breaks only.

Preliminary Steps

If you want to get started with the logstash-gelf support for WildFly, you will need to integrate logstash-gelf as a module within the server. It is possible to create the module manually, but it's easier to use the prepackaged binary. Please follow the steps to integrate the module:

Obtain the logging module from Maven or Github
Extract the module to <WildFly Home>/modules/system/layers/base.
Configure the logging subsystem by either adopting the changes in the XML files or by using the CLI

WildFly Logging configuration

XML Configuration:

<formatter name="JsonFormatter">
    <custom-formatter module="biz.paluch.logging" class="biz.paluch.logging.gelf.wildfly.WildFlyJsonFormatter">
        <properties>
            <property name="version" value="1.0" />
            <property name="facility" value="logstash-gelf" />
            <property name="fields" value="Time,Severity,ThreadName,SourceClassName,SourceMethodName,SourceSimpleClassName,LoggerName,NDC" />
            <property name="extractStackTrace" value="true" />
            <property name="filterStackTrace" value="true" />
            <property name="includeLogMessageParameters" value="true" />
            <property name="includeLocation" value="true" />
            <property name="mdcProfiling" value="true" />
            <property name="timestampPattern" value="yyyy-MM-dd HH:mm:ss,SSS" />
            <property name="additionalFields" value="fieldName1=fieldValue1,fieldName2=fieldValue2" />
            <property name="additionalFieldTypes" value="fieldName1=String,fieldName2=Double,fieldName3=Long" />
            <property name="mdcFields" value="mdcField1,mdcField2" />
            <property name="dynamicMdcFields" value="mdc.*,(mdc|MDC)fields" />
            <property name="includeFullMdc" value="true" />
        </properties>
    </custom-handler>
</formatter>

<file-handler name="JsonLog">
    <file relative-to="jboss.server.log.dir" path="server.json"/>
</file-handler>
...

<root-logger>
    <level name="INFO"/>
    <handlers>
        <handler name="FILE"/>
        <handler name="CONSOLE"/>
        <handler name="JsonLog"/>
    </handlers>
</root-logger>

CLI Configuration:

/subsystem=logging/custom-formatter=JsonFormatter/:add(module=biz.paluch.logging,class=biz.paluch.logging.gelf.wildfly.WildFlyJsonFormatter,properties={ \
       version="1.0", \
       facility="logstash-gelf", \
       fields="Time,Severity,ThreadName,SourceClassName,SourceMethodName,SourceSimpleClassName,LoggerName,NDC", \
       extractStackTrace=true, \
       filterStackTrace=true, \
       includeLogMessageParameters=true, \
       includeLocation=true, \
       mdcProfiling=true, \
       timestampPattern="yyyy-MM-dd HH:mm:ss,SSS", \
       additionalFields="fieldName1=fieldValue1,fieldName2=fieldValue2", \
       additionalFieldTypes="fieldName1=String,fieldName2=Double,fieldName3=Long", \
       mdcFields="mdcField1,mdcField2" \
       dynamicMdcFields="mdc.*,(mdc|MDC)fields" \
       dynamicMdcFieldTypes="my_field.*=String,business\..*\.field=double" \
       includeFullMdc=true \
})

/subsystem=logging/file-handler=JsonLog/:add(file={"relative-to"=>"jboss.server.log.dir", path=server.json}, \
        level=ALL,named-formatter=JsonFormatter)

/subsystem=logging/root-logger=ROOT/:add-handler(name=JsonLog)