Hello
When devices are too overloaded; we got lot of gap in the graph and it is not very nice.
To avoid that, i would like to implement the abberrant behavior detection provided by RRDTOOL.
Following http://oss.oetiker.ch/rrdtool/doc/rrdcreate.en.html, i defined a new RRA in ~xymon/server/etc/rrddefinitions.cfg
[test_rrd] RRA:AVERAGE:0.5:1:576 RRA:AVERAGE:0.5:6:576 RRA:AVERAGE:0.5:24:576 RRA:AVERAGE:0.5:288:576 # RRA:HWPREDICT:<array length>:<alpha>:<beta>:<period> # enable RRAs for aberrant detection, 1 day seasonal store for 5 day RRA:HWPREDICT:1440:0.1:0.0035:288 <------ new entry
i got a simple test script running every 5 mn:
sh-3.00$ cat test_rrd.sh
#!/bin/sh
HOSTNAME="test_device"
value=42
$BB $BBDISP "status+200d $HOSTNAME.test_rrd green date
test: $value"
i got ncv conf in xymonserver.cfg, the graph is working fine with this basic definition:
[test_rrd] TITLE test RRD YAXIS value DEF:a=test_rrd.rrd:test:AVERAGE LINE2:a#00CC00:test value GPRINT:a:LAST: \: %5.1lf (cur) GPRINT:a:MAX: \: %5.1lf (max) GPRINT:a:MIN: \: %3.1lf (min) GPRINT:a:AVERAGE: \: %5.1lf (avg)\n
now, if i add the def:
DEF:b=test_rrd.rrd:test:HWPREDICTthe graph is not showing anymore [even if i didn t use the new DEF on a LINE] :(
if i dump the rrd i can see the RRA :
sh-3.00$ rrdtool dump /opt/xymon/data/rrd/test_device/test_rrd.rrd |grep HW <cf> HWPREDICT </cf>
any idea is welcome.
Thanks.
regards Nico