Statistics-cumulative

Unbound
1

Hello to all, im following the http://www.unbound.net/documentation/howto_statistics.html doc to setup unbound stats using munin, in the doc is not clear if a have to enable statistics-cumulative or not, however in the munin plugin itself it indicated that statistics-cumulative are required.
I have run some tests and i got huge values like 663.67m bogus querys, i think if for the acumulative setting beacuse i just got 9.6k qps in the hits graph.
Can someone explain this?
regards

2

Hi Miguel,

Hello to all, im following the
http://www.unbound.net/documentation/howto_statistics.html doc to
setup unbound stats using munin, in the doc is not clear if a have
to enable statistics-cumulative or not, however in the munin plugin
itself it indicated that statistics-cumulative are required. I have
run some tests and i got huge values like 663.67m bogus querys, i
think if for the acumulative setting beacuse i just got 9.6k qps in
the hits graph.

663.68 mega or milli ?

600.000.000 qps is a bit high, yes. If you do unbound-control
stats_noreset you can peek at the statistics output. The output uses
plain numbers (without k,m, ...). Can you try this and see what
numbers are wrong there?

Best regards,
   Wouter

3

hi wouter, if i check in the munin graph there is 16.59m bogus querys, running the command line i got:

[root@dns1 ~]# unbound-control stats_noreset | grep bogus
num.answer.bogus=0
num.rrset.bogus=1
[root@dns1 ~]#

I think m is for mega, similar as k if for kilo right?
in the other hand just for test, i disable the cumulative setting and the graph show a huge fall in the numbers, take a look a this picture:

Im wondering what is the correct value, with or without cumulative ?

4

Hi everybody,

We are running the munin plugin without cumulative statistics and the numbers we get seem to be correct. We also have a cacti server (also based on rrdtool) so I have some previous experience with this exellent tool. If I my memory serves me right, m is for milli and M is for mega. Also, if you change anything that affects the data collection your graphs and numbers will be screwed up for a while since it is calculating average values.

Hope this is helpful.

// Anders

21 mar 2013 kl. 18:14 skrev “Miguel Miranda” <miguel.mirandag@gmail.com>:

5

thanks Anders, you are right, after one and a half day after i turned off cumulative statistics i got correct values,
thats all, thanks again.