SRX

Hi Guys,

I have an question about how to view current average utilization of CPU and memory on SRX3600.

 I know we can see through menu dashboard on Jweb, but it's weird sometimes is 100% cpu utlization (red color) and if we show through CLI: show chasiss routing-engine, it show sometimes idle is 98% and sometimes idle is 10%.

I want to see the average of utilization of CPU and memory, because it's fluctuating i can't  calculate the average.

any way to see the current average of cpu, memory utilization in SRX ?

Thank you Guys

Hi,

Have a read here:

http://forums.juniper.net/t5/SRX-Services-Gateway/SRX-220-Sudden-Very-High-Sustained-CPU-Utilization/td-p/151292

Hi,

I have read it. Thank you for the info.

What i want to know is, can we just show total average cpu utilization that curently active in SRX, because running :

"show chassis routing-engine" we can see the detail information about all cpu utilization but if we refresh (run command : show chassis routing-engine the result is change and the different can be significant.

for example : sometimes 90% idle and sometimes 10% idle. SRX is currently not processing traffics, because application that will be through SRX is still develop. When apps is ready, it will through and process by SRX.

I have difficulties to show the data about average of SRX system status (average CPU, memory utilization), I have to provide the current average utilization to customer, because they want that information and as i told above that the  command :"show chassis routing-engine" the result is always fluctuacting and changing signinficantly, so i confuse to know the average utilization).

Hi there, show chassis routing-engine shows the load avg, if you are on 11.3 or above, there is some handy SNMP for CPU load averages(below).  Otherwise you would have to use something like Solarwinds SNMP monitoring software to produce some graphs etc for you.  This may be nice to show the customer also.

http://kb.juniper.net/InfoCenter/index?page=content&id=KB23726

Run the following command to see these MIBs in the CLI on a SRX to get 1, 5 and 15 minute load averages:

root@srx> show snmp mib walk jnxOperatingEntry | match LoadAvg.9.1.0.0
jnxOperating1MinLoadAvg.9.1.0.0 = 18
jnxOperating5MinLoadAvg.9.1.0.0 = 12
jnxOperating15MinLoadAvg.9.1.0.0 = 8

Hi,

So we have to use SNMP monitoring tools like solarwind to see the average cpu utilzation?

Or any other way to see directly from SRX?

If we see load average from "show chassis routing-engine" : there are 1, 5, 15 minutes with it's value for each. 

Example in 5 minutes is  0.06 what is the meaning of 0.06 ? Is it mean that cpu utilization in 5 minutes is 6% ?

Hi there,

Yes and No is the answer.

Have a read here:

http://www.linuxjournal.com/article/9001

user@srx>show snmp mib walk jnxOperatingCPU

If you want a avg based over a few days or something I can see no other way but to use Solar Winds or the likes.

Hi,

What about my previous question : 

If we see load average from "show chassis routing-engine" : there are 1, 5, 15 minutes with it's value for each. 

Example in 5 minutes is  0.06 what is the meaning of 0.06 ? Is it mean that cpu utilization in 5 minutes is 6% ?

Refer to the article above:

"The load averages differ from CPU percentage in two significant ways: 1) load averages measure the trend in CPU utilization not only an instantaneous snapshot, as does percentage, and 2) load averages include all demand for the CPU not only how much was active at the time of measurement."

"If there are four CPUs on a machine and the reported one-minute load average is 4.00, the machine has been utilizing its processors perfectly for the last 60 seconds. This understanding can be extrapolated to the 5- and 15-minute averages."

Hi,

So load average is a total number of CPU utilization in SRX, ex : if it has 4 CPU on a machine and load average in 1 minute is 2.00 then means that in 60 seconds just using 2 cpu, right?

So if we want to make sure the CPU condition status wheter its using all cpu or just some cpu , we can see in load average, right?

and if we want to see cpu trend utilization in 30 minutes or 1 hour, has to using snmp monitoring tools, right?

How can we measure the cpu utilization is heavy load processing or full use in SRX CLI or WEB?

Thanks 

If the box has 4 CPUs and the 1 minute load average is 2.0, it means that the trend load average is around 50% across the cpus.  It narrowly focuses on what is actively demanding CPU time, where as the CPU % is the amount if time a process was found on the CPU during a sampling period.

If you wnt to monitor the load in real time during heavy throughput you could use the top- H from the comand shell.  The load averages are displayed in real time at the top of the screenshot below.

user@srx>start shell
% top -H

Hi,

based on the output capture :

- CPU states is : 76,4% , it means that active processing/cpu utilization in all cpu currently is 76,4% ?

- what about WCPU : cpu1 is 90.97%, cpu2 is 90,19%, and cpu3 is 90,19%. are those percents means total of idle or active processing ?

Thanks

The SRX uses a multicore processor.  The flowd_octeon thread usually runs at high percentage in order to make sure that all the packets received are processed faster.  This is why flowd reads high always, however it does not mean that system is busy.

The actual CPU usage for Control and Data Plane can be found here

For Routing Engine (Control Plane) CPU usage:

  show chassis routing-engine

For Packet Forwarding Engine (Data Plane) CPU usage:

  show chassis forwarding

  show security monitoring fpc 0

The summary line shows the percentage the cpu has spent in the various processor states. Notice, If you add them you will get 100%.  User+System+Nice+Interrupt+Idle = 100%.

Hi,

Thank you for your information...

Really helpfull.