[smokeping-users] Smokeping RTT Avgration matcher alert

[Jonny Graves]

Hello,

I'm trying to tune the Avgratio matcher alert. I'm currently monitoring DNS
latency for ~250 endpoints from ten slaves located around the globe and a
single master. The DNS probe is configured ping five times every five
minutes. RTTs across these slaves to various endpoints range from 200 us to
350 ms. I'm trying to create a single RTT alert that will only when there
is a sustained significant increase in RTT for a given endpoint. This has
proven difficult since the locations with the really low RTTs tend to be
really noisy regardless of how high I set thresholds. Here is my current
pattern:
+rttdetect
type = matcher
pattern =
Avgratio(historic=>60,current=>15,comparator=>'>',percentage=>350);
edgetrigger = yes
priority = 1
comment = test for round trip times

Any suggestions?
Also, it appears that whenever I tweak the thresholds to an alert and
reload the Smokeping configuration it appears that the alert won't start
until Smokeping has polled the amount times configured in the pattern as
opposed to using the preexisting historical data. Can someone confirm this
behavior is correct? Thanks in advance for any info or assistance that you
can offer.

Regards,
Jonny Graves
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.oetiker.ch/pipermail/smokeping-users/attachments/20161202/60db4019/attachment.html>