whats the best way?...15 switches

[Anil]

Hi

I have 15 switches, each with a different IP address that I need to monitor (i shall do so via icmpping and just see if they are up/down). I was thinking of creating a host group called "switches" for example, and then just adding 1 item (icmpping) for each swtich. However this doesn't seem very efficient and I was wondering if anybody had a better way of doing this.

[cooper]

You could create on template host with icmpping set as the only item and then create you 15 switch hosts based on that template. It wont save you creating all the hosts, but it will save you from adding icmpping to all of them. HTH

cooper.

[morgolis]

either way it's up to you how you set up your host, the template is a good idea to save you from having to set up each monitored item, but i suggest you use the simplecheck command: tcp,23
This actually forces a good TCP connect, not just a wimpy ICMP reply.
Only thing to keep in mind is that it doesn't disconnect cleanly. We found this out on our Redbacks when it would connect but no cleanly disconnect, it would just wait for the connection to time out. So set your time interval to 5 mins or so, that should give ample time to allow the TCP connection to time out on the router. Kinda depends on what type of switch you have and how it's managed. If they are Cisco Catalysts, i suggest the tcp,23 connect but if they are web managed, you can also do a tcp,80 check as well.

Also, on your trigger is suggest you set it up as such:
hostname:tcp,23.min(timeinterval x 3)
This will give you a "hold down" timer for false checks, by setting whatever time interval you have multiplied by 3, it in essence says "Only throw this trigger if i have failed making the connect three times." Saves alot of false alarms.