Oskar, I'd like to see the output more human-usable. The square column on the time-based stats isn't terribly useful unless you're bringing it into a spreadsheet or somesuch (or do very long division and square roots in your head well ;-).
For an Excel (sorry, I am Satan) spreadsheet that grabs stats via localhost HTTP and does some nice graphs, see: CHK at Lcy3pdVUduruyDMRa-R0BsN~sT4RAwI,sR-7TLJYBsiyJREWafGq3A I assume this isn't compatible with the new stats output, but it works like a champ with the days-old build, and I'll get it updated when the new formats settle down. Two stats I'd like to see the system provide directly: QueryRejections (inboundAggregateRequests - inboundAggregateRequestsRejected), and some kind of "store hit rate", which I think would be the % of requests that were served from the local store. I believe that stores are generally full, but if that's not correct, a stat indicating how much free space there is in the store might be interesting. -glenn > -----Original Message----- > From: devl-admin at freenetproject.org > [mailto:devl-admin at freenetproject.org]On Behalf Of Oskar Sandberg > Sent: Thursday, February 07, 2002 3:06 AM > To: devl at freenetproject.org > Subject: Re: [freenet-devl] average hop time > > > On Thu, Feb 07, 2002 at 12:21:41AM -0500, Tavin Cole wrote: > > What kinds of average hop times are people seeing since Oskar fixed the > > diagnostics bugs? I'm seeing values in the 2.0-3.5 range, on hawk and > > on another, transient, node. > > > > We may be able to significantly reduce the hopTimeExpected and > > hopTimeDeviation values in the conf, which should cut back the long > > waits on some requests. > > It should be noted, again, that the deviation on those results is the > deviation of the mean value over the number of hops made per request, > not the deviation of the per hop value. Roughly, the per hop devitation > would be that times sqrt(<average number of htl on requests>), so maybe > we should add a field to look at htl. > > Does anybody have any wishes regarding other statistical data we should > keep? Min and max as well as frequency diagrams should be possible > (though the latter would take a bit of memory) - I am less certain about > things like medians and quartiles since they seem difficult > (implossible?) to aggregate. The stat files are unfortunately not very > forwards compabitle, though I consider them a disposable comodity. > > > -- > > Oskar Sandberg > oskar at freenetproject.org > > _______________________________________________ > Devl mailing list > Devl at freenetproject.org > http://lists.freenetproject.org/mailman/listinfo/devl _______________________________________________ Devl mailing list Devl at freenetproject.org http://lists.freenetproject.org/mailman/listinfo/devl
