Skip to content

Latest commit

 

History

History
76 lines (52 loc) · 2.55 KB

README.md

File metadata and controls

76 lines (52 loc) · 2.55 KB
Raw

Tests Homebrew Release

jvm-mon

Console based JVM monitoring - when you just want to SSH into a server and see what's going on.

jvm-top lets you monitor your JVM server applications from the terminal.

Screenshot

New Version

Release: 1.0-ea1

  • Rewritten in Go
  • Single executable file
  • Can monitor applications on Java 8 and above
  • Does not require an existing JDK

How it works:

  • jvm-mon executable comes bundled with a Java agent jar
  • On startup it extracts the agent to a temp directory
  • It attaches to the JVM you want to monitor
  • Loads agent into running JVM to collect metrics
  • Agent and app establish a socket connection to send metrics

Install

Requirement: a JDK8 on the server and JAVA_HOME environment variable pointing to it. It won't work with just a JRE.

MacOS

brew install jvm-mon

Linux/MacOS

  1. Download the release and extract
  2. Set JAVA_HOME environment variable: export JAVA_HOME=/path/to/your/jdk8
  3. Execute ./bin/jvm-mon from extracted directory

Usage

  • Select a JVM process and press Enter to monitor it
  • Press q or Ctrl+C to exit
  • Press Del or Backspace to kill a process

What is available

Currently it shows:

  • List of running JVM processes
  • Cpu and GC load
  • Heap size and usage
  • Top threads with cpu usage

Building from source

To build locally run ./gradlew installDist. Then go to ./build/install/jvm-mon/ and run ./bin/jvm-mon.

To develop you will need npm on your machine and then run ./gradlew npmDeps once to get dependencies.

How does it work?

jvm-mon is a Kotlin application based on these awesome libraries:

  • blessed-contrib terminal dashboard library in JavaScript
  • J2V8 Java Bindings for V8 JavaScript engine and Node.js
  • jvmtop Java monitoring for the command-line

The way it works is:

  1. The Kotlin app starts a Node.js engine in-process
  2. Node.js loads a script with all the widgets
  3. The script calls back into Kotlin to get metrics