I am trying to calibrate the new CPU counters like ssCpuRawUser (50),
ssCpuRawNice (51) and ssCpuRawSystem (52) without much success.

Whilst mrtg is reading the registers the movement recorded in the log is
often 0.  The few entries shown here were caused by a demanding task that
vmstat averaged at about 10% utilisation however mrtg only inferred 1% and
2%.

root@terabite:/var/www/mrtg# head -10 fw_cpu.log
1365584642 965303 759068
1365584642 0 0 0 0
1365584343 0 0 0 0
1365584100 0 0 0 0
1365583800 0 0 0 0
1365583500 0 2 1 3
1365583200 1 2 1 3     <--- this entry analysed
1365582900 0 0 1 2
1365582600 0 0 0 0
1365582300 0 0 0 0

Extracts taken from the logs and analysed were ...

timestamp    raw-I    raw-O    delta-I   delta-O log-I log-O
1365582900   964380   757278   56        50      0     0
1365583200   964660   757880   280       602     1     2

The config I am using is ...

#
# Firewall CPU
#
Target[FW_cpu]:
.1.3.6.1.4.1.2021.11.50.0&.1.3.6.1.4.1.2021.11.52.0:[email protected] +
.1.3.6.1.4.1.2021.11.51.0&.1.3.6.1.4.1.2021.11.55.0:[email protected]
Options[FW_cpu]: growright, nopercent
MaxBytes[FW_cpu]: 400
Title[FW_cpu]: Firewall CPU Load
PageTop[FW_cpu]: <H1>Firewall CPU Load</H1>
Unscaled[FW_cpu]: ymwd
ShortLegend[FW_cpu]: %
YLegend[FW_cpu]: CPU Utilisation
LegendI[FW_cpu]: User
LegendO[FW_cpu]: System
#

Is this an interplay of bits versus bytes affecting scaling?  I think this
machine has 2 cores but I expected that would inflate the number rather than
deflate it as I am seeing.

Cheers, Frank

_______________________________________________
mrtg mailing list
[email protected]
https://lists.oetiker.ch/cgi-bin/listinfo/mrtg

Reply via email to