Collectd aggregation-cpu-sum wrong value?

[Branislav Kirilov]

This is tcpdump from freenas server posting to graphite

Code:

tcpdump -i lagg0 -Avv dst host ......... | grep aggregation | grep cpu-idle
tcpdump: listening on lagg0, link-type EN10MB (Ethernet), capture size 65535 bytes
............aggregation-cpu-sum.cpu-idle 1969.409672 1484318185
............aggregation-cpu-sum.cpu-idle 1969.845383 1484318195
............aggregation-cpu-sum.cpu-idle 1982.818131 1484318205
............aggregation-cpu-sum.cpu-idle 1980.089518 1484318215
............aggregation-cpu-sum.cpu-idle 1983.622353 1484318225
............aggregation-cpu-sum.cpu-idle 3944.586813 1484318235
............aggregation-cpu-sum.cpu-idle 1980.201147 1484318255
............aggregation-cpu-sum.cpu-idle 1982.254528 1484318265
............aggregation-cpu-sum.cpu-idle 1976.795328 1484318275
............aggregation-cpu-sum.cpu-idle 1979.525364 1484318285

So the values freenas posts are 1969.409672. In top and in the internal reporting section i see cpu-idle is 93%.
How is aggregation-cpu-sum.cpu-idle being calculated?
I tried /cpus which is 16, but result is still not correct. This applies to all aggregation stats.

 

[dlavigne]

Were you able to find an answer to this?

 

[Branislav Kirilov]

Unfortunately no. I have no idea how this value is generated.