On Tue, May 10, 2011 at 10:46:33AM -0400, Nick Anderson wrote:
>On Tue, 2011-05-10 at 16:26 +0200, [email protected] wrote:
>> Cfengine's monitoring is very different and superior to anything else I've
>> seen. Cf notices when things are abnormal for that particular host. Normal
>> monitoring solutions use centralized thresholds that are then manually
>> changed for each exceptional host.
>>
>> For example a load of 5 or less might be considered normal for all hosts
>> except some who are set higher. This is the traditional way. Cfengine does
>> away with this. The monitor daemon statically analyzes the host and
>> determines what is normal. No guessing. No generalizing.
>
>So is there a way to see what cf-engine thinks is a "normal" range?
Poke around in /var/cfengine/reports/*.{q,E-sigma,distr}.
.q is the last raw value for that timeinterval , in the 2nd column, and
(I think) "how long ago in hours" the data was collected in column 1.
(I'm not sure on this--need to dig into it more).
For .E-sigma, column 1 is the same as for .q. Column 2 might be the
average, and column 3 is the stdev. (again, not really sure).
The .distr file is actually a histogram, with column 1 as the "bucket"
value (x-axis), and column 2 as the frequency count for that bucket.
Not everything that has a file appears to actually be populated (based
on cf-monitord running in promiscuous mode...).
I also don't know what cfenv-{average,now,stddev} are for (yet).
--
Jesse Becker
NHGRI Linux support (Digicon Contractor)
_______________________________________________
Help-cfengine mailing list
[email protected]
https://cfengine.org/mailman/listinfo/help-cfengine
