Recommended Cisco Switch Thresholds

I'm monitoring a number of Cisco switches / interfaces. I know how to set up alert thresholds, I just don't know what's important to monitor.

Do I need to alert on all of the parameters, or are there a few key Node/Interface statistics that I should focus on? What do you alert on?

Thank you for any recommendations.

Find more posts tagged with

threshold

Alerts

parameter

Accepted answers

All comments

almay66

I was wondering the same thing. What do other users monitor? Is there a baseline that we can follow?

Miles.cass

I do not have a great answer for you but I can tell you this, every network is different and you will see wildly different results from one network to another as far as baseline numbers. That beeing said what I would recomend is establish a few criteria that you konw is important to keeping your orginzation running without complaints.

An example may be internet bandwidth, you konw that if it reaches 95% utlization users will not have a plesant experiance. Once you establish what items are key to be monitoring I would then look at some of the data Orion has already collected and figure an approximate average utlization during a "Normal" time on your network and jot those numbers down. You can take that number and create an alert and have Orion alert you if that number is greatly exceeded for a certian period of time. Over time you may need to adjust these numbers that better suit your needs but it should get you started.

As far as I konw there are no hard and fast baseline number everyone uses its all about what works best for you.

Some items you may be intrested in watching:

- Node CPU utilization

- Memory Utlization

- Interface bandwidth on routers

- Response times on devices

Etc.

Hope this was of some help.

MagnAxiom

Some of the things that we monitor in addition to the list Miles.cass listed:

- Interface Errors (Alert if you cross a threshold of over 50 packet errors per hour)

- Fan Status (A fan dying in a switch can be bad stuff... better to know about it as soon as possible.)

- Temperature - (Again, a switch overheating is bad. Tracking environmental stats can extend the life of your gear)

- Interface Status (Is the interface of interests Up/Down, pretty basic)

- Reboot (Watch the last boot time, if it changes, the switch as rebooted... so alert about it)

Hope this helps.

MagnAxiom