What does the "MAX CPU" measure in Operations Monitor dashboard really mean?

[mwallman]

Hi all,

This might sound like really simple question but I want to understand it more deeply.

In Operations Monitor app there is a measure called Max CPU in the Performance sheet.

Every Hour, or 10 Minutes it presents what the CPU was like at that time.

But how does it actually capture the Max CPU?

If the CPU reached 94% for example for a millisecond during a time period, does it classify that as the Max CPU for that Hour or 10minute period when it happened?

[Tyler_Waterfall]

Hi @mwallman - thanks for the question.

Max CPU comes from the Engine service in Sense and is polled every X minutes (set in the QMC > Engine settings). It represents the CPU of the Engine service only, not the entire server (or other services).

As described (very sparsely!) in the online help (HERE), 

Qlik Sense Engine Service (QES)

Each entry (that is, row) in the Performance log corresponds to a snapshot (that is, a number of measurements) of the performance of the QES at the given point in time...

---

• CPULoad: A measurement of the load on the CPU on which the QES runs at the given point in time.

---

This doesn't answer your question completely, though, as you are wondering about sampling or averaging vs point in time.

As @StefanBackstrand pointed out a long time ago HERE, "This value is the average of the % CPU load recorded during the measure interval. As stated in the QVS Reference Manual, page 205, for SR1."

In Operations Monitor, Max CPU just takes the max value it finds for CPULoad within the given time period across all (hostname) data.