arthas-mcp
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| ARTHAS_WS_URL | No | Arthas WebSocket endpoint (HTTP API URL is derived) | ws://127.0.0.1:8563/ws |
| ARTHAS_HTTP_URL | No | Alternative: give the HTTP API URL instead | |
| ARTHAS_MAX_OUTPUT | No | Per-response character cap before truncation | 16000 |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| arthas_execA | Execute an Arthas command and return the result. For sync commands (dashboard, thread, jvm, sc, sm, jad, ognl, etc.) returns result directly. For streaming commands (watch, trace, stack, monitor) runs in background and returns a job ID + log file path. |
| arthas_dashboardC | Get a comprehensive dashboard showing thread stats, memory usage, GC info, and runtime data. |
| arthas_threadC | Get thread information. Shows thread stack traces, CPU usage, and thread states. |
| arthas_jvmC | Get JVM information including memory usage, GC stats, runtime info. |
| arthas_memoryC | Get detailed memory information. |
| arthas_scC | Search for loaded classes. |
| arthas_smC | Search for methods in loaded classes. |
| arthas_jadC | Decompile a loaded class. |
| arthas_watchA | Watch method execution. Runs in background; waits up to awaitMs for completion and returns results inline if done, otherwise returns a job ID. Use arthas_read_log to check output later, arthas_stop_job to stop. |
| arthas_traceA | Trace method execution path and timing. Runs in background; waits up to awaitMs for completion and returns results inline if done, otherwise returns a job ID. |
| arthas_stackA | Get the call stack of a method. Runs in background; waits up to awaitMs for completion and returns results inline if done, otherwise returns a job ID. |
| arthas_monitorA | Monitor method invocation statistics. Runs in background; waits up to awaitMs for completion and returns results inline if done, otherwise returns a job ID. |
| arthas_ognlC | Execute OGNL expressions. |
| arthas_classloaderC | Show classloader hierarchy and info. |
| arthas_getstaticC | Get static field values. |
| arthas_heapdumpB | Create a heap dump file. |
| arthas_profilerC | Start/stop async profiler for CPU or memory profiling. |
| arthas_ttC | Time Tunnel: Record method invocations for replay. Record mode (record=true) is ASYNC. |
| arthas_versionB | Get Arthas version information. |
| arthas_helpB | Get help for Arthas commands. |
| arthas_cookbookA | Get a proven step-by-step Arthas command recipe for a common production symptom (high CPU, memory leak, slow requests, worker pool exhaustion, deadlocks, connection pool, Spring inspection, ...). START HERE when diagnosing a symptom instead of composing commands from scratch. Call without arguments to list topics. |
| arthas_async_jobsA | List all async background jobs (watch/trace/stack/monitor). Shows job ID, command, status, log file path, and elapsed time. |
| arthas_read_logA | Read the output log of an async job. Use this to check results from watch/trace/stack/monitor commands. Returns the tail by default to keep responses small. |
| arthas_stop_jobA | Stop a running async job (watch/trace/stack/monitor). |
| arthas_set_configC | Update Arthas connection configuration. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 25 tools
Each tool has a distinctly described purpose. While some tools (e.g., watch, trace, stack, monitor) overlap in monitoring methods, their descriptions clearly differentiate them (watch for watching parameters/return, trace for timing, stack for call stack, monitor for stats). No ambiguity remains.
All tool names follow the consistent pattern `arthas_<verb_or_noun>`, using snake_case throughout. The prefix is uniform, making the set predictable and easy to navigate.
With 25 tools, the server covers the essential Arthas diagnostic commands without being overwhelming. Each tool serves a specific purpose, and the count balances comprehensiveness with usability.
The tool set covers all major Arthas capabilities: class inspection, method tracing, thread analysis, JVM info, profiling, heap dump, OGNL, async job management, and a cookbook. There are no obvious gaps for common diagnostic tasks.