01-07-2010 10:40 AM - edited 03-06-2019 09:11 AM
What would you guys say is an acceprable level for normal switch operation? Not for the below switch but just in general.
LVL3-LDRP-48-POE
CPU utilization for five seconds: 8%/0%; one minute: 8%; five minutes: 8%
PID Runtime(ms) Invoked uSecs 5Sec 1Min 5Min TTY Process
109 97114377 43517427 2231 0.15% 0.14% 0.15% 0 HRPC qos request
127 1743836777 236669849 7368 1.59% 1.38% 1.40% 0 Marvell wk-a Pow
144 697456938 802606357 868 1.11% 0.90% 0.95% 0 Spanning Tree
#show proc cpu | e 0.0001-07-2010 10:51 AM
Hi Mike
Lower the number, better it is (always).. but i wouldnt be alarmed to see CPU utilizations from 20-50 % ... On older switches like 2900 XL & 3500 XL I think the CPU is always more than 50 %, even in normal circumstances....
With new switches which are highend, you can see much lower CPU instances.. We normally configure the network management system to throw an alert say after 70 % ... again, we are looking at a longer duration for CPU spikes, and not a 1 sec interval ! It again is based on what type of traffic your router/switch runs, and if the device was properly positioned for that particular task..
I would baseline the CPU utlization of the devices, and see what the acceptable levels are for devices,... report the baseline cpu for a month, and then configure NMS systems accordingly... anything over 50 % i would definitely have a look at what process is taking up the CPU, especially on core switches and highed edge components..
Hope this helps.. all the best..
Raj
01-07-2010 12:53 PM
Thanks Raj. Thats exactly what i want to do is a baseline and go from there. Just wasnt sure on some of the older stuff what acceptable cpu usage was. Thanks again.
Find answers to your questions by entering keywords or phrases in the Search bar above. New here? Use these resources to familiarize yourself with the community: