[j-nsp] High CPU and interrupts

Alex alex.arseniev at gmail.com
Wed Jul 1 16:13:01 EDT 2009

JUNOS calculates kernel and user CPU separately so what you want to monitor 
is idle CPU.
Idle CPU close to 0 (in single digits) over prolonged period is worth 
AFAIK, the is no Juniper OID for 5min averaged CPU but you can always write 
an event script and populate jnxUtil MIB with numbers from "show chassis 
routing-engine" (note that only master RE CPU is displayed so for backup RE 
CPU you have to run the script on backup RE as well and somehow get the data 
from backup RE to master RE).

----- Original Message ----- 
From: "harbor235" <harbor235 at gmail.com>
To: <juniper-nsp at puck.nether.net>
Sent: Wednesday, July 01, 2009 3:13 PM
Subject: [j-nsp] High CPU and interrupts

>I am trying to configure a monitoring system to alarm using thresholds for
> CPU utilization and
> interrupts. Junos has separation in the control and data planes so i would
> assume brief spikes
> in CPU should not be a problem. However, sustained high COU should be an
> indicator for a investigation
> and troubleshooting. Cisco has avgBusy5 OID that captures the cpu overage
> over a 5 minute interval,
> does Junos have something similar? I would think that a sustained CPU
> utilization of >90% should warrant
> investigation, is there anything similar for interrupts and interface
> stats that anyone is looking at?
> Mike
> _______________________________________________
> juniper-nsp mailing list juniper-nsp at puck.nether.net
> https://puck.nether.net/mailman/listinfo/juniper-nsp

More information about the juniper-nsp mailing list