JMeter Result Analysis Plugin

A Maven plugin that parses JMeter result XML files and generates detailed reports with charts

Can be used in combination with the JMeter Maven Plugin that is developed by the same authors

Features

• Text and HTML output of certain statistics (minimum, maximum, average, standard deviation, quantiles) for response duration and response size
• Output is rendered with Freemarker and can be customized
• Chart containing request duration and average of all requests
• CSV file containing durations of all response (by URL)
• CSV file containing sizes of all responses (by URL)
• Statistics and charts can be generated per request group. Request groups are defined by URL patterns.
• Download of resources from remote systems for the JMeter execution interval

Build status

Buildhive CI:

Help

Run the HelpMojo of this plugin: mvn jmeter-analysis:help -Ddetail=true -Dgoal=analyze

Release Notes

See Release Notes in the Wiki.

Usage Example

<plugin>
  <groupId>com.lazerycode.jmeter</groupId>
  <artifactId>jmeter-analysis-maven-plugin</artifactId>
  <executions>
    <execution>
      <goals>
        <goal>analyze</goal>
      </goals>
      <configuration>
        <!--
        An AntPath-Style pattern matching a JMeter XML result file to analyze. Must be a fully qualified path.
        File may be GZiped, must end in .gz then.

        Default: not set.
        Required.
        -->
        <source>${project.build.directory}/**/*.jtl</source>

        <!--
        directory where to store analysis result files.

        Default: ${project.build.directory}
        Required.
        -->
        <targetDirectory>${project.build.directory}/results</targetDirectory>

        <!--
        Request groups as a mapping from "group name" to "ant pattern".
        A request uri that matches an ant pattern will be associated with the group name.
        Request details, charts and CSV files are generated per requestGroup.

        The order is honored, a sample will be added to the first matching pattern. So it's possible
        to define various patterns and one catch all pattern.

        If not set, the threadgroup name of the request will be used.

        Default: not set.
        -->
        <requestGroups>
          <pages>/page/**</pages>
          <binaries>/binary/**</binaries>
        </requestGroups>
        <requestGroups>
          <requestGroup>
            <name>pages</name>
            <pattern>/page/**</pattern>
          </requestGroup>
          <requestGroup>
            <name>binaries</name>
            <pattern>/binary/**</pattern>
          </requestGroup>
        </requestGroups>

        <!--
        If set to true, additional files "<category>-sizes.csv" and "<category>-durations.csv" will be stored.
        These files contain detailed information for response size and response durations for every URI.

        Default: true
        -->
        <generateCSVs>true</generateCSVs>

        <!--
        If set to true, additional chart files "<category>-durations.png" will be created.

        Default: true
        -->
        <generateCharts>true</generateCharts>

        <!--
        If set to true, the directory structure relative to source will be preserved during output.

        Default: false
        -->
        <preserveDirectories>false</preserveDirectories>

        <!--
        Set<String> of sample names that should be processed when analysing a results file.

        Default: sample, httpSample
        -->
        <sampleNames>
          <sampleName>sample</sampleName>
          <sampleName>httpSample</sampleName>
        </sampleNames>

        <!--
        If set to true will process all files found by the pattern defined in <source>.
        If set to false (the default) it will only process the first file found.

        *SETTING TO TRUE IS NOT RECOMMENDED*
        It can substantially impact performance, if you do this you do it at your own risk!

        Default: false
        -->
        <processAllFilesFound>false</processAllFilesFound>

        <!--
        Template directory where custom freemarker templates are stored.
        Freemarker templates are used for all generated output. (CSV files, HTML files, console output)
        Templates must be stored in one of the following three subfolders of the templateDirectory:

        csv, html, text

        The entry template must be called "main.ftl".

        For example,
        templateDirectory/text/main.ftl will be used for generating the console output.

        Default: not set.
        -->
        <templateDirectory>${project.basedir}/src/main/resources/</templateDirectory>

        <!--
        Mapping from resource URL to file name. Every resource will be downloaded and stored in 'targetDirectory'
        with the given filename. Tokens "_FROM_" and "_TO_" can be used as placeholders. These placeholders will
        be replaced by timestamps of execution interval (formatted as ISO8601, e.g. '20111216T145509+0100').

        Default: not set.
        -->
        <remoteResources>
          <property>
            <name>http://yourhost/path?from=_FROM_&amp;to=_TO_</name>
            <value>my_resource.txt</value>
          </property>
        </remoteResources>

		<!--
		Specify custom date format for resources not supporting ISO8601.
		
		Default IOS8601
		-->
		<remoteResourcesFromUntilDateFormat>HH:mm_yyyyMMdd</remoteResourcesFromUntilDateFormat>
		
        <!--
        List of writers that handle all output of the plugin.
       Defaults to:
       * com.lazerycode.jmeter.analyzer.writer.ChartWriter (generates detailed charts as PNGs),
       * com.lazerycode.jmeter.analyzer.writer.DetailsToCsvWriter (generates CSV files for every request group),
       * com.lazerycode.jmeter.analyzer.writer.DetailsToHtmlWriter (generates HTML files for every request group),
       * com.lazerycode.jmeter.analyzer.writer.HtmlWriter (generates an HTML overview file),
       * com.lazerycode.jmeter.analyzer.writer.SummaryTextToFileWriter (generates a TXT overview file),
       * com.lazerycode.jmeter.analyzer.writer.SummaryTextToStdOutWriter (generates overview output to stdout)

       If one of those should be deactivated or a new com.lazerycode.jmeter.analyzer.writer.Writer implementation should be added,
       all desired writers need to be configured!
        -->
        <!--<writers>-->
          <!--<com.lazerycode.jmeter.analyzer.writer.SummaryTextToStdOutWriter/>-->
          <!--<com.lazerycode.jmeter.analyzer.writer.SummaryTextToFileWriter/>-->
          <!--<com.lazerycode.jmeter.analyzer.writer.HtmlWriter/>-->
          <!--<com.lazerycode.jmeter.analyzer.writer.DetailsToCsvWriter/>-->
          <!--<com.lazerycode.jmeter.analyzer.writer.DetailsToHtmlWriter/>-->
          <!--<com.lazerycode.jmeter.analyzer.writer.ChartWriter/>-->
        <!--</writers>-->

      </configuration>
    </execution>
  </executions>
</plugin>

Example Output

An analysis summary text output looks like this:

time: 20111216T145509+0100 - 20111216T145539+0100
requests:             36049
requests per second:  1201
total duration:       30
response duration (ms)
  min:                0
  average:            0
  max:                1352
  standard deviation: 7
  quantiles (ms)
       10%        0
       20%        0
       30%        0
       40%        0
       50%        1
       60%        1
       70%        1
       80%        1
       90%        1
       99%        6
     99.9%       19
    100.0%     1352 (max. value)
response size (bytes)
  total:              750210890
  min:                20480
  average:            20810
  max:                53890
  standard deviation: 3308
response status codes
  200:               36049 (100%)

Credits

Part of the development of this plugin is sponsored by CoreMedia