Date: Fri, 29 Mar 2024 06:13:54 +0000 (UTC) Message-ID: <540394704.5625.1711692834524@ip-10-0-1-161.ec2.internal> Subject: Exported From Confluence MIME-Version: 1.0 Content-Type: multipart/related; boundary="----=_Part_5624_557633737.1711692834522" ------=_Part_5624_557633737.1711692834522 Content-Type: text/html; charset=UTF-8 Content-Transfer-Encoding: quoted-printable Content-Location: file:///C:/exported.html
You may notice that the graphing of CPU Total and CPU by Process=
(e.g. Workload top 10) don't match on multi-CPU or multi-core Windows serv=
ers. For example, you may see that your total CPU averages 50% but your top=
10 shows 3 processes all at 70% CPU.
This difference is related to the way Windows tracks its performance count=
ers at the process level. The maximum CPU usage is always 100% at the CPU t=
otal level. At the per-process level, the total will be 100% x the number o=
f cores. So on a 2-CPU dual-core server, your per-process workload may go a=
s high as 400%.
Keeping this in mind, you can quickly divide per-process CPU workload by t=
he number of cores to get a true understanding of how much of the total CPU=
a single process is consuming instead of seeing how much of a single core =
it is consuming.