Hi Ed: How are you gathering the statistics and how is that being fed to gmetric? I wonder if you could merge the data collection with your gmetric cronjob. Also, how frequently is your cronjob run?
Cheers, Bernard On Wed, Jul 1, 2009 at 12:08 PM, Ed Greenberg<[email protected]> wrote: > I have a statistic that I'm gathering once per minute. I run a cron job > that feeds gmetric. (This is with ganglia 3.1.1 on Centos. (I compiled > it myself since it's not available by RPM.) > > The data item is a counter and my gmetric command line looks like this: > gmetric -n Searches -v $COUNT -t uint32 -s positive > > What I get for my minute by minute data input is this (fragment): > > <database> > <!-- 2009-07-01 11:21:00 PDT / 1246472460 --> > <row><v> 0.0000000000e+00 </v></row> > <!-- 2009-07-01 11:21:15 PDT / 1246472475 --> > <row><v> 2.2611111111e+01 </v></row> > <!-- 2009-07-01 11:21:30 PDT / 1246472490 --> > <row><v> 4.5222222222e+00 </v></row> > <!-- 2009-07-01 11:21:45 PDT / 1246472505 --> > <row><v> 0.0000000000e+00 </v></row> > <!-- 2009-07-01 11:22:00 PDT / 1246472520 --> > <row><v> 2.2648888889e+01 </v></row> > <!-- 2009-07-01 11:22:15 PDT / 1246472535 --> > <row><v> 3.4844444444e+00 </v></row> > <!-- 2009-07-01 11:22:30 PDT / 1246472550 --> > <row><v> 0.0000000000e+00 </v></row> > <!-- 2009-07-01 11:22:45 PDT / 1246472565 --> > <row><v> 0.0000000000e+00 </v></row> > <!-- 2009-07-01 11:23:00 PDT / 1246472580 --> > <row><v> 7.8600000000e+00 </v></row> > <!-- 2009-07-01 11:23:15 PDT / 1246472595 --> > <row><v> 1.8340000000e+01 </v></row> > <!-- 2009-07-01 11:23:30 PDT / 1246472610 --> > <row><v> 0.0000000000e+00 </v></row> > <!-- 2009-07-01 11:23:45 PDT / 1246472625 --> > <row><v> 0.0000000000e+00 </v></row> > <!-- 2009-07-01 11:24:00 PDT / 1246472640 --> > <row><v> 8.6857142857e+00 </v></row> > <!-- 2009-07-01 11:24:15 PDT / 1246472655 --> > <row><v> 2.1714285714e+01 </v></row> > <!-- 2009-07-01 11:24:30 PDT / 1246472670 --> > <row><v> 0.0000000000e+00 </v></row> > <!-- 2009-07-01 11:24:45 PDT / 1246472685 --> > <row><v> 0.0000000000e+00 </v></row> > <!-- 2009-07-01 11:25:00 PDT / 1246472700 --> > <row><v> 4.8470588235e+00 </v></row> > <!-- 2009-07-01 11:25:15 PDT / 1246472715 --> > <row><v> 2.2619607843e+01 </v></row> > <!-- 2009-07-01 11:25:30 PDT / 1246472730 --> > <row><v> 0.0000000000e+00 </v></row> > <!-- 2009-07-01 11:25:45 PDT / 1246472745 --> > <row><v> 0.0000000000e+00 </v></row> > <!-- 2009-07-01 11:26:00 PDT / 1246472760 --> > <row><v> 0.0000000000e+00 </v></row> > <!-- 2009-07-01 11:26:15 PDT / 1246472775 --> > <row><v> 2.5312500000e+01 </v></row> > <!-- 2009-07-01 11:26:30 PDT / 1246472790 --> > <row><v> 8.4375000000e-01 </v></row> > <!-- 2009-07-01 11:26:45 PDT / 1246472805 --> > <row><v> 8.4375000000e-01 </v></row> > > Note the zeroes. > > Here are my day and hour graph: http://www.greenberg.org/graphs > > So my hour graph shows individual points with spaces between them, but > they don't correspond very well to the data I'm seeing. My day graph > seems to reflect the magnitude of the data in the non-zeo cells. > > I'd really appreciate an explanation of what's going on, and what the > best way is to clean this up. > > Thanks, > </edg> > > Ed Greenberg > > ------------------------------------------------------------------------------ > _______________________________________________ > Ganglia-general mailing list > [email protected] > https://lists.sourceforge.net/lists/listinfo/ganglia-general > ------------------------------------------------------------------------------ _______________________________________________ Ganglia-general mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/ganglia-general
