Monitoring Part 2

[michael4702]

Ok now I know what I need to monitor, what thresholds should I look at for my filters?

Is there any best practice guides?

How do I know if my switch is under to much load, has to many errors on any one port?

Many thanks.

Mike.

[sean]

In a healthy network, you should not see too many errors, if any. You might see a lot of discards, but those can occur from traffic trying to exceed a rate-limit. Errors are of more concern. Best practice would be to baseline your network for about 2 weeks to establish what is normal. One done, this will give you an idea of what your network is like and you can work from there trying to tweak and clean it. Hope this helps.

[michael4702]

And what would be the max cpu ussage you would want to see on any given switch?

Many thanks.

Mike.

[devang_etcom]

use following command to see the cpu ussage

#show process cpu

hope this will help

rate all help full post

regards

Devang

[sean]

The max cpu would depend on the switch. If you have an access layer switch, such as a 2950, then an average of 25% or less would be resonable assuming you aren't doing anything facy with this. You will see bursts, but the average is what is important for long term. The burst could help you possible track down some virus or worm traffic. Hope this helps.