Yes. I watch NTP offsets from ntpstat.rrd with the following in
analysis.cfg:
HOST=%^(host1|host2)
# alert if NTP offset exceeds 10ms
DS ntp ntpstat.rrd:offsetms >10 COLOR=red "TEXT=NTP offset &Vms is
higher than &Ums"
DS ntp ntpstat.rrd:offsetms >5 COLOR=yellow "TEXT=NTP offset &Vms
is higher than &Ums"
DS ntp ntpstat.rrd:offsetms <=5 COLOR=green "TEXT=NTP offset &Vms
is within +/-&Ums"
DS ntp ntpstat.rrd:offsetms <-10 COLOR=red "TEXT=NTP offset &Vms is
lower than &Lms"
DS ntp ntpstat.rrd:offsetms <-5 COLOR=yellow "TEXT=NTP offset &Vms
is lower than &Lms"
Also, I do a similar thing for DNS latency, where the graphs are
automatically generated when I add a "dns" test configuration to hosts.cfg:
# alert if lookups take too long (200ms or 50ms)
DS dns tcp.dns.rrd:sec >0.05 COLOR=yellow "TEXT=Response time &V
exceeds &U seconds"
DS dns tcp.dns.rrd:sec >0.2 COLOR=red "TEXT=Response time &V
exceeds &U seconds"
These go into my SLA-type reports to management.
Cheers
Jeremy
On 13 December 2012 01:55, Kevin King <user-4c6c31ddcab2@xymon.invalid> wrote:
Anyone doing any alerting on graphed rrd data? Just looking for some
sample of what you might be doing.****
** **
-Kevin****
** **