Oddly, I was thinking about something similar to this this afternoon. One thing that occurred to me was to alter rrdtool, adding the ability to support a 'sparse' format, where instead of inferring the time of a row by it's position in the file, as I assume rrdtool normally does, it writes a more traditional timestamp for each row.
This would be a loss on frequently updated metrics, but a win on infrequently updated ones. Then, the decision would have to be made as to which type to use in each case. Rick Cobb wrote: > I'm still fairly ignorant about the whole rrd interface, but doesn't > counter support also require us to have more flexibility in declaring > the RRA(s) per RRD? Of course, basic counter support would still count > to one instead of many, but if the strategy for determining "counter- > ness" is to provide this information in meta-data attributes other > than "type", perhaps counting to many is called for. > > -- ReC > On Aug 29, 2008, at 6:31 AM, Brad Nicholes wrote: > >>>>> On 8/29/2008 at 2:34 AM, in message >> <[EMAIL PROTECTED] >> , <[EMAIL PROTECTED]> wrote: >> >>> >>> >>> I notice there is only one set of RRAs defined in gmetad.conf >>> >>> We've identified that most of our RRD storage is used by short term >>> data >>> (10 seconds between consolidated data points). >>> >>> However, for some metrics, this short term data has no benefit - e.g. >>> the mem_total metric, which would only change if the machine is >>> upgraded. >>> >>> One solution that I can think of is to classify each metric (either >>> within gmond or within gmetad), and then have a different `RRAs' >>> parameter in gmetad for each class of metric. Is anyone already >>> implementing such a strategy or an alternative solution that may >>> address >>> this problem? >>> >> I don't know of anybody that is working on this. It sounds like an >> interesting idea. Is the purpose just to reduce the size of >> the .rrd files? How would differing RRAs affect the web front end? >> Would the web front end have to change the way that it queries the >> RRDs when it creates the graphs? >> >> Brad >> >> >> ------------------------------------------------------------------------- >> This SF.Net email is sponsored by the Moblin Your Move Developer's >> challenge >> Build the coolest Linux based applications with Moblin SDK & win >> great prizes >> Grand prize is a trip for two to an Open Source event anywhere in >> the world >> http://moblin-contest.org/redirect.php?banner_id=100&url=/ >> _______________________________________________ >> Ganglia-developers mailing list >> Ganglia-developers@lists.sourceforge.net >> https://lists.sourceforge.net/lists/listinfo/ganglia-developers > > > ------------------------------------------------------------------------- > This SF.Net email is sponsored by the Moblin Your Move Developer's challenge > Build the coolest Linux based applications with Moblin SDK & win great prizes > Grand prize is a trip for two to an Open Source event anywhere in the world > http://moblin-contest.org/redirect.php?banner_id=100&url=/ > _______________________________________________ > Ganglia-developers mailing list > Ganglia-developers@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/ganglia-developers ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/ _______________________________________________ Ganglia-developers mailing list Ganglia-developers@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ganglia-developers
