Open Kedesai opened 8 years ago
We will review this issue ASAP thank you to report.
Ok I did the function in Graphite to scale at 10 and the values are close matching scale(.infrastructure.xxxx.xxx.xxxxx.esx.xxxxxxxxhost.mem.consumed_kiloBytes_average,1000) now it shows scale like 12.0G, 14.0G, 18.0G etc this should be workable. I can closely match the host memory consumed from what I see in the Vsphere client. The ballooning may be an issue since that one dipped.
Hi @Kedesai , There is a known but related to cpu.usage.average values in the native "statsfeeder" tool.
You can see here https://labs.vmware.com/flings/statsfeeder#comment-539192 for the issue and the "workaround".
In other way your problem could be related to this detected bug https://github.com/SYNAXON/GraphiteReceiver/issues/34 , we are working in a fix right now.
Are you working with "only_one_sample_x_period = true" ?
Hi Toni-moreno,
I am getting the same problem with CPU utilization as noted by #16. I ran the statsfeeder and collected the stats in .csv file and looks like proper percentage but when Graphitereceiver posts to Graphite the decimals are in the wrong place for eg 1500.0. Also Ballooning (mem.vmmemctl_kiloBytes_average) goes below 0 % not sure how that can be that can be the case. The mem.consumed_kiloBytes_average shows incorrect % on y axis. Can you help?