>>> On 5/2/2007 at 4:14 PM, in message <[EMAIL PROTECTED]>, <[EMAIL PROTECTED]> wrote: > Revision: 773 > http://svn.sourceforge.net/ganglia/?rev=773&view=rev > Author: bnicholes > Date: 2007-05-02 15:14:28 -0700 (Wed, 02 May 2007) > > Log Message: > ----------- > rework the interface for the metric modules so that a single module can > gather multiple metrics. This also allows the metric names to be different > from the module name. Use the -m commandline parameter to discover the names > of the metrics that the metric module supports. > > Modified Paths: > -------------- > trunk/monitor-core/gmond/gmond.c > trunk/monitor-core/gmond/metric.h > trunk/monitor-core/gmond/modules/conf.d/example.conf > trunk/monitor-core/gmond/modules/example/mod_example.c >
I just committed the above patch that allows a module to support multiple metrics. The purpose for this is to be able to support metric gathering such as per-CPU metrics on a multi-processor system or individual disk metrics for a machine that contains multiple disks. I am currently experimenting with a module that will discover how many CPUs are on the box and dynamically create individual metrics per CPU. If anybody has some code that will gather per-CPU utilization metrics, please pass it along. Otherwise I will try to derive something from the CPU utilization metric code that already exists. Brad
