Cpu usage on atlas

Hello ,

On altlas for CPU which matrics we should concern ?

i gone through with mongo Doc and they suggest to check normalize System CPU and normalize process CPU .

on my servers normalize System CPU shows almost ident between 10 -20 % while normalize process CPU is upto 60% . should it is concern?

Regards,

Daksh

Hi @Dasharath_Dixit,

Thank you for your question! System CPU and Normalized Process CPU are both useful in estimating processor requirements. Of the two, Normalized Process CPU may be easier to understand as it already factors in the number of CPU cores and presents the data as a percentage between 0 and 100%.

Regarding your normalized process CPU, our default threshold on this metric is currently 95%. I don’t believe 60% is a concern unless you see that it continues to grow and nears 95% or you see that the maximum value regularly exceeds 95%.

Thanks,
Frank

2 Likes

Hello @Frank_Sun ,

Thanks for the response!

got it . but mostly normalize system CPU is between 10-20 % and normalize process CPU is much higher .

second for both metrics it is mention that , the value is divide by number of core . so the metrics on GUI is already shown the value after the calculation based on VCPU ?

Regards
Dash

Hi @Dasharath_Dixit,

Ah sorry I misread your original question. Yes, both the normalized system CPU and normalized process CPU metrics do divide by the number of vCPUs. If you are seeing normalized system CPU lower than your normalized process CPU, could you please open a ticket and include your group ID so we can further investigate?

Thanks,
Frank

1 Like

hello @Frank,

Yeah i already raised a ticket to mongo and got the update as below


Please note that the **Normalized Process CPU** displays the following information:

* `user` displays the percentage of time that the CPU spent servicing the **MongoDB process** , scaled to a range of 0-100% by dividing by the number of CPU cores.
* `kernel` displays the percentage of time the CPU spent servicing operating system calls for the **MongoDB process** , scaled to a range of 0-100% by dividing by the number of CPU cores.

While the **Normalized System CPU** to monitor CPU usage as this displays the CPU usage of **all processes on the node** , scaled to a range of 0-100% by dividing by the number of CPU cores.

Generally, for best clarity, we recommend that you use the **Normalized System CPU** as this displays the CPU usage of **all processes on the node** , scaled to a range of 0-100% by dividing by the number of CPU cores that the cluster has.

I normally noticed that my all cluster server has normalize system CPU around 10 -25% while nomalize process CPU is higher . which i guess normal but i really dont understand why there are multiple CPU related matrics and what exactly difference between them …

regards
Dash

Hi @Dasharath_Dixit,

but i really dont understand why there are multiple CPU related matrics and what exactly difference between them …

You can see definition of each metric within the metrics page of your cluster(s) and select the info icon as detailed here in this post. It is a bit of an older post but I believe you can see the definition for each in the same way described in it.

Regards,
Jason