I'm trying to get the metric named "CPU|Usage (MHz)" in vROPS via API.
From the API I know that the "key" is:
description=Average CPU usage in MHZ
So the UI for Jul 14 05:26AM actually gives me 17,521 as shown below.
But the API return for the same time gives me 2173.466552734375 which is obviously not nearly close.
api return: datetime(2019-07-14 05:26:59.999000-03:00) value=(2173.466552734375)
The VM has 8 vCPUs so I could multiply that value by 8 which gives me 2173 * 8 = 17.384 which is closer but not accurate all the times.
Doing some more digging I figured out that it only "represents" the value of the first CPU of the VM:
cpu:0|usagemhz_average 2019-07-14 05:26:59.999000-03:00 2173.466552734375
cpu:1|usagemhz_average 2019-07-14 05:26:59.999000-03:00 2191.60009765625
cpu:2|usagemhz_average 2019-07-14 05:26:59.999000-03:00 2192.39990234375
cpu:3|usagemhz_average 2019-07-14 05:26:59.999000-03:00 2192.0
cpu:4|usagemhz_average 2019-07-14 05:26:59.999000-03:00 2190.86669921875
cpu:5|usagemhz_average 2019-07-14 05:26:59.999000-03:00 2193.666748046875
cpu:6|usagemhz_average 2019-07-14 05:26:59.999000-03:00 2190.0
cpu:7|usagemhz_average 2019-07-14 05:26:59.999000-03:00 2182.533447265625
Summing all the cpu mh values give me a more precise value: 17506.533447265625 but not the same.
So.. question is... Is this expected? Is there another way of fetching this?
Additional notes: I could also try using (CPU|Demand, cpu|demandmhz) key but as it turns out this doesn't work to reserved VMs as it only shows the max configured value.
Can anyone help?
My vROPS Version