jmeter-mcp-server 2.0.0

Advanced MCP server for JMeter performance test analysis with 14 comprehensive tools

JMeter MCP - Complete Guide

Package: jmeter-mcp v1.0.0 | Status: ✅ Ready for Publication | Date: January 27, 2026

---

🚀 Quick Start

What is This?

A stdio-based MCP server for analyzing JMeter performance test logs. Works with Autogen and any LLM via MCP protocol.

Installation & Usage

# No installation needed - run directly
pipx run jmeter-mcp

# Or install once
pip install jmeter-mcp
jmeter-mcp

# Use with Autogen (zero-install)
from autogen.agentchat.contrib.mcp.agent import MCPAgent
agent = MCPAgent(
    name="analyst",
    mcp_server_config={
        "type": "stdio",
        "command": "pipx",
        "args": ["run", "jmeter-mcp"],
    },
)

---

📦 Package Contents

Source Code

jmeter_mcp/
├── __init__.py           Package metadata
└── cli.py               MCP server (231 lines, 3 tools)

Build Artifacts

dist/
├── jmeter_mcp-1.0.0-py3-none-any.whl    Ready to install
└── jmeter_mcp-1.0.0.tar.gz              Ready to upload

Configuration

pyproject.toml           Package config (modern PEP 517/518)
LICENSE                  MIT License
verify.py                Verification script (8/8 checks)

---

🔧 Available Tools

1. analyze_jmeter_logs

Analyzes JMeter performance test logs and returns key metrics.

Input: log_data (CSV format)
Output: Metrics (total requests, failures, error rate, response times)

2. detect_log_format

Auto-detects log format with confidence score.

Input: log_data (first few lines)
Output: Format type and confidence (jmeter_jtl, performance_csv, etc.)

3. get_supported_formats

Lists all supported format types.

Input: None
Output: Array of supported formats with descriptions

---

📝 Publishing to PyPI (4 Steps)

Step 1: Update Metadata (2 min)

Edit pyproject.toml and replace:

authors = [
    {name = "YOUR NAME", email = "[e-mail]"}
]

[project.urls]
Homepage = "https://github.com/YOUR_USERNAME/jmeter-mcp"
Repository = "https://github.com/YOUR_USERNAME/jmeter-mcp.git"

Step 2: Create PyPI Account (5 min)

1. Visit: https://pypi.org/account/register/
2. Create account (free)
3. Verify email
4. Generate API token
5. Save token

Step 3: Publish (1 min)

python -m twine upload dist/*

# When prompted:
# Username: __token__
# Password: [paste your token]

Step 4: Verify (5-10 min)

# Wait for PyPI to index
pipx install jmeter-mcp
pipx run jmeter-mcp 2>&1

# Output: "Starting JMeter MCP Server (stdio transport)"

---

🤖 Autogen Integration

Basic Setup (Zero Installation)

from autogen import UserProxyAgent
from autogen.agentchat.contrib.mcp.agent import MCPAgent

mcp_config = {
    "type": "stdio",
    "command": "pipx",
    "args": ["run", "jmeter-mcp"],
}

analyst = MCPAgent(
    name="performance_analyst",
    system_message="Analyze JMeter logs using available tools",
    mcp_server_config=mcp_config,
    llm_config={"config_list": [{"model": "gpt-4"}]},
)

user = UserProxyAgent(name="user", human_input_mode="TERMINATE")
user.initiate_chat(analyst, message="Analyze this JMeter test...")

Multi-Agent Collaboration

# Create team of agents
analyst = MCPAgent(name="analyst", ...)
reporter = AssistantAgent(name="reporter", ...)
manager = UserProxyAgent(name="manager")

# Group chat
group_chat = GroupChat(agents=[analyst, reporter, manager])
chat_manager = GroupChatManager(group_chat=group_chat)

manager.initiate_chat(chat_manager, message="Analyze and report...")

---

✅ Verification & Testing

Run Verification

python verify.py

# Output: 8/8 checks passed ✅

Checks Include:

• Package structure correct
• pyproject.toml valid
• Documentation complete
• CLI imports working
• MCP tools defined
• Stateless design verified
• Logging to stderr
• Build artifacts present

Test Locally

# With pip
pip install -e .
jmeter-mcp 2>&1

# With pipx
pipx install dist/jmeter_mcp-1.0.0-py3-none-any.whl --force
pipx run jmeter-mcp 2>&1

---

🎯 Key Features

✅ Stdio-only - Pure JSON-RPC 2.0, no HTTP
✅ Stateless - No file I/O, parameter-driven
✅ Zero-install - Works via pipx run jmeter-mcp
✅ Production-grade - Error handling, input validation
✅ Autogen-ready - Direct integration support
✅ Well-documented - Complete guides
✅ Free - MIT license, PyPI hosting

---

🔧 Commands Reference

# Verify everything is ready
python verify.py

# Test locally with pip
pip install -e .

# Test locally with pipx
pipx install dist/jmeter_mcp-1.0.0-py3-none-any.whl --force
pipx run jmeter-mcp 2>&1

# Publish to PyPI
python -m twine upload dist/*

# Install from PyPI (after publishing)
pip install jmeter-mcp

---

🆘 Troubleshooting

Issue	Solution
"Module not found: twine"	pip install twine build
"401 Unauthorized"	Generate new token at https://pypi.org/account/
Server doesn't start	Run jmeter-mcp 2>&1 to see stderr errors
"Package already exists"	Update version to 1.0.1 in pyproject.toml
"No such command: jmeter-mcp"	Reinstall: pip install --force-reinstall jmeter-mcp

---

📊 Project Stats

Metric	Value
Package Size (wheel)	7.7 KB
Package Size (source)	452 KB
Python Support	3.9, 3.10, 3.11, 3.12, 3.13
Core Dependencies	2 (mcp, fastmcp)
MCP Tools	3
Verification	8/8 ✅
Time to Publish	~20 minutes

---

🔐 Constraints & Features

Constraint	Status	Details
Stdio-only MCP	✅	JSON-RPC 2.0 over stdin/stdout
No HTTP/FastAPI	✅	Pure stdio, no web framework
Stateless operation	✅	Parameter-driven, no session state
No file I/O	✅	All data via JSON parameters
Logs to stderr	✅	stream=sys.stderr configured
Works with pipx run	✅	Entry point registered & tested
Free publication	✅	PyPI hosts packages free

---

📁 Folder Structure

jmeter-mcp-package/
├── jmeter_mcp/                      Python package
│   ├── __init__.py
│   └── cli.py
├── dist/                            Build artifacts
│   ├── jmeter_mcp-1.0.0-py3-none-any.whl
│   └── jmeter_mcp-1.0.0.tar.gz
├── pyproject.toml                   Modern package config
├── LICENSE                          MIT License
├── verify.py                        Verification script
└── README.md                        This complete guide

---

🎬 Publishing Workflow

1. Update pyproject.toml metadata (2 min)
   ↓
2. Create PyPI account (5 min)
   ↓
3. Generate & save API token (2 min)
   ↓
4. Run: python -m twine upload dist/* (1 min)
   ↓
5. Wait 5-10 minutes for PyPI indexing
   ↓
6. Test: pipx install jmeter-mcp
   ↓
✅ PUBLISHED!

---

🎁 Use Cases

Performance Analysis

# Analyze JMeter test results
pipx run jmeter-mcp

With Claude/Autogen

# Analyze logs with AI assistance
agent.chat("Analyze this JMeter result and find issues")

CI/CD Integration

# In GitHub Actions or Jenkins
pipx run jmeter-mcp < test_results.csv

---

📞 Quick Links

Link	Purpose
https://pypi.org	Python Package Index
https://pypi.org/account/register/	Create PyPI account
https://modelcontextprotocol.io	MCP specification
https://microsoft.github.io/autogen/	Autogen documentation

---

✨ Next Steps

1. Review this guide
2. Update pyproject.toml with your name/email
3. Verify everything: python verify.py
4. Create free PyPI account
5. Publish: python -m twine upload dist/*
6. Test: pipx install jmeter-mcp

---

📜 License

MIT License - Free to use, modify, and distribute

---

Status: ✅ Production Ready | Created: January 27, 2026

Ready to publish? Update pyproject.toml and run: python -m twine upload dist/*