“Power Meets Simplicity in Just Three Clicks”

MightyMeter

• Introduction
• Requirements
• Installation of the Leader
• Installation of the Workers
• Configuring Your JMX File
• Running Tests
• Grafana Dashboards
• Querying the Database
• Additional Commands
• Generating Data

Introduction

MightyMeter is a powerful performance testing solution that leverages InfluxDB, Grafana, and JMeter to support both local and distributed testing modes. Here’s a closer look at its two operational modes:

1. Local CLI Mode

In local testing, it’s recommended to run JMeter in CLI (Command Line Interface) mode instead of the GUI. The CLI mode is more efficient for executing tests and conserving resources, which makes it ideal for performance evaluation. While JMeter’s GUI mode is highly useful for creating and modifying .jmx test files, the CLI mode ensures stable, reliable execution when running actual tests.

2. Distributed Testing Mode

MightyMeter facilitates a straightforward setup for distributed testing with multiple JMeter servers. In this configuration, a central Leader node (or master node) coordinates multiple Worker nodes (or slave nodes). The Leader node distributes the test configurations to each Worker, and as the test runs, each Worker sends its results back to the Leader. This mode is particularly valuable for large-scale tests and scenarios requiring load generation from multiple locations or servers.

Key Features in Both Modes:

• Result Persistence: Test results are stored in an InfluxDB database, providing a durable and structured record of performance metrics.
• Real-Time Monitoring and Visualization: MightyMeter integrates with Grafana dashboards to enable real-time test monitoring. These visualizations pull data from InfluxDB and provide insights into various test statistics. Additional dashboards allow you to compare metrics across different application versions, user loads, and other test conditions.
• Comprehensive Reporting: Test reports are generated in both text and HTML formats and include APDEX (Application Performance Index) ratings. Reports are accessible via an HTTP server powered by NGinx, making it easy to review and share performance summaries.

In summary, MightyMeter offers a flexible testing environment suited to both local and distributed load testing needs, with robust support for result persistence, real-time insights, and detailed reporting.

Requirements

To set up the system, ensure you have the following:

• Docker installed
• Download the latest release of MightyMeter
• Extract the downloaded file and grant execution permissions for the scripts in the bin folder
• It’s recommended to add the following paths to your shell’s environment PATH variable:

MIGHTY_METER_HOME="PATH_TO_HOME"

chmod -R +x ${MIGHTY_METER_HOME}/bin

export PATH="$MIGHTY_METER_HOME/bin/leader:$MIGHTY_METER_HOME/bin/worker:$PATH"

Installation of the Leader

For the machine intended as the Leader and MightyMeter downloaded, run the following shell script:

mm.leader.host$ mm-leader-install

Services on the Leader will be set up as follows:

• Nginx is listening on port 80.
• Grafana is accessible at http://localhost/grafana (set up as a reverse proxy through Nginx, default login is admin/admin).
• InfluxDB is running in a container. Refer to Querying the Database for instructions on opening the CLI.
• JMeter leader container is created at the start of a test and removed once the test is complete.
• In addition to Grafana dashboards, you can view APDEX JMeter reports at http://localhost/jmeter/reports in HTML format and download them as jtl files. The JTL format is used by JMeter to store performance test results, typically in CSV files. Each JTL file logs data gathered during a test, including metrics such as response times, status codes, sample names, and other performance indicators.

Port can be customized via the variable _NGINX_PORT in the .defaults file. Component versions are specified in the .versions file:

_INFLUXDB_VERSION=2.7.10
_GRAFANA_VERSION=11.3.0
_NGINX_VERSION=1.27.2
_JMETER_VERSION=5.6.3
_JMETER_OPEN_JDK_VERSION=21-oraclelinux8

You are now ready to use MightyMeter in CLI mode from your local machine! Check out this example to get started.

Installation of the Workers

To set up distributed testing, first install the Leader on a designated machine. Next, install MightyMeter on each Worker machine. For example, on each Worker machine with hostnames such as mm.worker.host1, mm.worker.host2, download MightyMeter and execute the following shell script on each Worker:

mm.worker.host1$ mm-worker-install --host mm.worker.host1

mm.worker.host2$ mm-worker-install --host mm.worker.host2

Don't install the worker nodes on the same machine as the leader—it defeats the purpose. For single-machine setups, use MightyMeter in local CLI mode instead.

Configuring Your JMX File

To persist results in InfluxDB and enable real-time monitoring in Grafana, include the MightyMeter listener in your JMX file, as shown in the jmx file example.jmx.

Here's how it looks in the JMeter GUI if you open the file:

You can only modify the following:

• The eventTags variable: set this to your chosen value (it will appear in the tags field within the InfluxDB events measurement). This value is for informational purposes only.
• Add custom tags in the format TAG_name=value.

Running Tests

Distributed testing

To start a distributed test, go to the leader machine and run the shell script mm-leader-run-tests with options like:

mm.leader.host$ mm-leader-run-tests \
--duration 3600 \
--num-threads 20 \
--ramp-up 60 \
--leader-host mm.leader.host \
--worker-hosts mm.worker.host1,mm.worker.host2 \
--env Production \
--jmx-file /path/to/jmx-file \
--props-file /path/to/props-file \
--app My App \
--app-version 1.0.0 \
--test-name PerformanceTest

Options:

• duration: test duration in seconds.
• num-threads: threads per Worker (total = num-threads * Worker count).
• ramp-up: ramp-up time for threads in seconds.
• leader-host: hostname of the Leader.
• worker-hosts: comma-separated Worker hostnames.
• env: test environment (used in InfluxDB and Grafana filters).
• app: application name.
• app-version: tested app version.
• jmx-file: JMeter configuration file (with listener included).
• props-file: additional properties for JMeter.

This command creates an ephemeral Docker container to start a JMeter client (Leader) and coordinates with active Workers. Monitor tests in real-time on Grafana at http://localhost/grafana (default login: admin/admin).

Local CLI Mode

localhost$ mm-cli-mode-run-tests \
--duration 20 \
--num-threads 10 \
--ramp-up 1 \
--env dev \
--jmx-file /path/to/jmx-file \
--props-file /path/to/props-file \
--app myapp \
--app-version 1.0.0 \
--test-name anotherTest

Options:

• duration: test duration in seconds.
• num-threads: threads per Worker (total = num-threads * Worker count).
• ramp-up: ramp-up time for threads in seconds.
• env: test environment (used in InfluxDB and Grafana filters).
• app: application name.
• app-version: tested app version.
• jmx-file: JMeter configuration file (with listener included).
• props-file: additional properties for JMeter.

This command creates an ephemeral Docker container to start a JMeter that executes the test. Monitor tests in real-time on Grafana at http://localhost/grafana (default login: admin/admin).

Grafana Dashboards

MightyMeter offers a range of dashboards to visualize both real-time and historical data:

Here are a few examples:

• For real-time monitoring:

  

• To compare two different versions of the same application under identical conditions:

  

• To compare the performance of the same application version under varying user loads:

  

Querying the Database

Access the InfluxDB console on the Leader with mm-leader-influxdb-console. Familiarity with InfluxDB terms like measurements, series, tags, and fields is helpful. The default database mm contains statistics and events. events logs test start/stop times; statistics holds transaction metrics.

Example:

mm.leader.host$ mm-leader-influxdb-console

> show DATABASES
> use "mighty-meter"
> select * from events
> select * from statistics
> quit

Additional Commands

• Uninstall MightyMeter: Run uninstall shell scripts on the Leader and each Worker

  mm.worker.leader$ mm-leader-uninstall

  mm.worker.host1$ mm-worker-uninstall

  mm.worker.host2$ mm-worker-uninstall

• Access Worker JMeter Logs:

  mm.worker.host1$ mm-worker-bash
  root@container# tail -f /var/log/jmeter/jmeter.log

• Access Leader JMeter Logs:

Go to http://localhost/jmeter/logs/

• Stop a running Test:

  mm.leader.host$ mm-leader-stop-test

---

Generating Data

Generating data becomes straightforward and highly efficient with tools like java-fun for any data structure and json-values for JSON.

To start generating data, simply place the JAR file containing the generators into JMeter's ext-libraries folder. For detailed instructions, refer to the java-fun project readme.

Please note that the necessary dependencies, json-values and java-fun, are already included in JMeter's lib-libraries folder.