[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: More statistics



On Feb 8, 2008 4:31 PM, Jeroen van Meeuwen <kanarip kanarip com> wrote:
> How about smolt only submitting the hardware data if the user chooses to
> do so, and otherwise just getting it's uuid in the database, instead of
> letting the user disable it entirely?


Smolt right now isn't a heartbeat... it doesn't ping out often enough
to be useful for doing moving averages, and on top of that i don't
know  if it makes since for live instances.

What i want if I can get it is a heartbeat which starts at boot up and
then pings a url at 6 hour intervals. Each live spin would ping a
specific url, and all harddrive installed systems would ping a common
arch specific url.  We do a moving average of the aggregate pings and
we'll get a much better idea of the number of nat'd systems. And if we
do it right I can even estimate a NAT'd system multiplier based on the
statistics of intervals between pings over time from the same ip
address...and at no point do I expose any individual ip addresses in
the resulting output.

I'm doing this sort of stuff now using the mirrorlist hits, but i'm
pretty sure yum-updates is used by 1% of the base.

-jef


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]