Node status via SNMP +

Question

As a new Orion user I have a couple of questions, which may or may not be totally obvious to you hardened veterans

First up I have a few nodes which are not pingable, but which are SNMP manageable. Orion shows the node status as down, but shows the interfaces as up, and collects stats correctly. Is there a way to get Orion to use the SNMP info to determine the node's status, rather than simply assuming that because it can't be pinged it must be down.

Secondly I've been experimenting with alerts, with some success. What I'd like to be able to do is to trigger an emailed alarm when a certain condition has held true for a certain amount of time. Specifically I'm not too worried about a brief peak in utilisation on a fairly busy WAN link, but if it becomes sustained (10 minutes maybe) I want to know. Likewise certain nodes may miss a single ping, but if they drop 2 or more I'd want to know. Is there any way to achieve this, rather than simply alerting on the first trigger?

Regards,

Richard

ttundra277 · Answer

I'm curious is if anyone has found a workaround for status polling using SNMP rather than ICMP.  It's easier said than done to convince a customer or vendor to allow ICMP requests through their firewall or on a public IP address when their security policies prohibit such traffic.  This should be considered by Orion and added as a feature to their applciation, as we can't provide certain stats on some devices to our customers, which really impacts the applications value.

mjhamilton · Answer

Welcome to the world of Orion (or Onion as we lovingly call it at work!!)..

You are a bit scuppered ref pinging the device.. SNMP is using port 162 (give or take 1) to monitor the services on the device such as disk, cpu and memory etc... however uptime and availability for the device is done purely by good old fashion pinging.... you will have to open the ICMP port on your firewall to allow this if you want to get uptime for devices and display their up/down status.... you should be able to create an explicit set of rules so that ICMP is only available for the Orion server to the monitored devices so it limits any security impacts.

As for missing pings etc... you can set the polling service to register a device as down only after a set number of polls have past.. I think this is what you are after.... you can only make this change globally so it will affect all monitored devices on the system... if you want to mix and match polling types you will need to set-up separate polling server instances to do this.... you will find this setting somewhere in the Orion Options area in the pull down menus of the console.. not 100% exactly where cos I am at home and cannot access at the moment..

Hope this is of some help.. let me know otherwise if I can help

Martin J. Hamilton... lost in a sea of snmp...

Network_Guru · Answer

Welcome to the forum Richard.

Orion uses ICMP to determine node status. There is currently no other way to make that Red dot Green.
As for delayed alerting, it is not supported per se, but you can just extend the statistics polling interval for the interface you are monitoring. The average utilization must exceed the set threshold for the full duration of the polling period.
Most of my interface stats are polled every 15 minutes, which works quite well.

As for nodes missing a single ping, you will never get an alert for just one missed ping.
There is a "fast polling timer" which pings a device multiple times for a set period of time before the node is marked as down.
See this post for more details;

www2.solarwinds.net/.../topic.asp

PS. Isaac - I have not tried Bawls, but I love the ThinkGeek site. Tons of cool stuff!

-=Cheers=-
          NG