tda

TDA (Thread Dump Analyzer) ships with an MCP Server mode that lets AI-assisted tooling analyze Java thread dumps and heap information in a headless setup. The server exposes TDA’s analysis capabilities so you can feed it dumps from automated pipelines or IDE workflows and receive structured insights, deadlock information, and thread state data that can be consumed by external agents or AI assistants. Use it to perform batch parsing of thread dumps, run JSON-based dumps (experimental), and leverage features like virtual thread analysis and carrier thread relationships in a headless context.

To use the MCP server, run the Java command to start the TDA server (e.g., java -Xmx512m -jar tda.jar) and connect your MCP-enabled tools or scripts to the running instance. The server will parse provided dumps and return analysis results compatible with MCP workflows, enabling AI models or automation scripts to extract meaningful diagnostics and feeding them into your AI-assisted development processes.

Prerequisites:

• Java 17 or newer installed on your system
• Access to the TDA distribution (tda.jar) or a built-from-source artifact

Installation steps:

1. Ensure Java is installed

   • macOS/Linux: java -version
   • Windows: java -version in Command Prompt

2. Obtain the TDA JAR (tda.jar) from the Releases page or build from source

3. Start the MCP server (headless analysis):

   bash java -Xmx512m -jar tda.jar

4. (Optional) If integrating with CI/CD or automation, point your MCP client/tools to the running server endpoint and use the MCP protocol to exchange analysis results.

Notes and tips:

• Memory: adjust the -Xmx flag according to the size of the dumps you expect to process (e.g., -Xmx1g for large analyses).
• JSON-based thread dumps are experimental; behavior and detail levels may vary, and some fields may be missing compared to textual dumps.
• If you’re integrating with AI tools, you can leverage MCP’s headless mode to feed dumps and receive structured analysis payloads suitable for downstream models.
• This is a Java-based server; no Node.js, Python, or Docker prerequisites are required unless you choose an alternative deployment method.
• If you encounter issues running the jar, verify that the Java runtime matches the project’s supported versions and that the working directory contains tda.jar.