[naemon-users] Nudging (max concurrent service checks)

Andreas Ericsson ageric79 at gmail.com
Mon Jan 11 14:03:14 CET 2016


On 2015-11-24 13:23, VERVAET Frederik (ITS/IIS) wrote:
> Hi,
>
> Naemon differs in behaviour from Nagios since instead of a random
> nudge interval it uses the retry count.
>
> In the logs the nudge X seconds is always 0 (the log message does not
> reflect the actual chosen interval) -> Trivial fix.
>
> Why was this behaviour chosen ? The semi random interval in Nagios
> flattened out peaks. In this case of the retry interval is the same
> (standard interval) it will not flatten out the peaks in load.
>

Nagios doesn't have a semi-random interval. It would always end up
scheduling all checks at the start of a new active period in the
timeperiod if there were any inactive periods, effectively causing a
cpu-saving intention of scheduling certain checks only during workhours
to clump up at 07:00 in the morning (or whenever your workhours period
started).

The "nudge one interval and see if we can run the check then" maintains
the original, random, distribution, causing checks to stay spread the
same way they did at startup.

/Andreas


More information about the Naemon-users mailing list