Grid Performance Monitoring System

Links:

[Project Page]
[Downloads]
[Tracker]
[Feedback (mail)]

Documentation:

[Quickstart guide]

Overview

GPM system is an actively developed distributed monitoring system for clusters or Grid nodes. Construction of system is based on a hierarchical design. The system can monitor cluster hosts and parallel jobs, running on these hosts.

The main feature of the monitoring system is the processing of clients requests. For example, you can request the list of hosts with free physical memory greater than 1GB (see the simple client: 9_3_findNodes).

Request output format is very flexible (for example: HTML, XML, other string formats).
Output expression is build from basic blocks: numbers, constants, metric values. All expressions may accept the arguments which are stored in the stack of arguments. Parameters can be specified in the request statically or passed by client or created dynamically during the request processing.
There are two types of expressions: simple expressions (condition; value; metric values; metric names) and composite expressions (output sequence of expressions; output a expressions with the dynamic set of parameters).

New! The first version of grid service released for Globus Toolkit 4.0.3. For more information see Quickstart guide

New! The first version of job information viewer released. For more information see Monitoring system demo section.

Monitoring system demo

The last version of the monitoring system is running on the some hosts of cluster of University of Nizhniy Novgorod.

Support

Your questions, comments and suggestions are welcome!
Email to:
gpm-system-help (at) lists.sourceforge.net

Also you can use: