SNMP + Item Anchoring Extensions (Tips & tricks)

by acaballero , Thursday, September 26, 2013, 06:36 (3859 days ago)
edited by acaballero, Thursday, September 26, 2013, 06:44

SNMP extension requirements and considerations

Please note that this extension is pretty barebones and is intented for people who knows about SNMP monitoring. You should document yourself on SNMP before attempting to use this extension on SMTPing, because you'll get lost easily.

This extension will let you specify an item as an SNMP node of a certain type. Right now it supports SNMP v1 and v2 (v3 will be added later on) and it comes with the Net-SNMP MIBs, which will be handy for monitoring Linux servers.

SNMP stuff NOT to miss

• If you have installed Net-SNMP thru CygWin or MinGW, it is for sure you've added the MIBDIRS variable in your system environment. Be warned that the package installer will override this variable and set it to its own MIB library location.
  
  You should backup the next registry value before installing the package:
  
  HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Session Manager\Environment\MIBDIRS
  
  Now, if you want SMTPing to use your MIBs library, you will need to, once installed, go to your system properties and change the MIBDIRS variable to the path where your MIBs reside.
  
  
• After installing the package, you may need to close/open your session or reboot your computer. That's because the MIBDIRS environment variable is going to be added by the installer.
  
  If you do a SNMP walk from an SMTPing node and you get the numeric or ISO OIDs, then you will need to reboot the computer, because the MIB reader needs to detect the MIBDIRS variable from the environment to load all the MIBs properly.
  
  
• If you rebooted the system and still can't get the OIDs on an SNMP walk from SMTPing, please drop us a support request here at the forum.

What the Item Anchoring extension is for

Due to the nature of the results given by SNMP agents (integers, percents, bytes, allocation units, etc.), you shouldn't mix up different types of results within the same group, because the group charts will be all messy if you have a regular ping node (measured in response MS), a TCP port checking node (measured as response in MS), a SNMP Free disk space node (measured in allocation units and converted to bytes) and a CPU activity node (measured in a percentile integer).

So, the correct usage of the SNMP nodes is pretty straight: you must create groups for distinct nodes with the same types of SNMP items inside (only CPU, only Disk, only interface, etc.) Then use the Item Anchoring extension to create a relationship between all of them.

Example:

• You are monitoring two servers, A and B.
  
  
• You create a group with two standard ping items, one for each server. You name them "Ping A" and "Ping B".
  
  
• You create an SNMP group for CPU monitoring with two SNMP nodes, one for each server.
  
  
• You create an SNMP group for Disk monitoring with two SNMP nodes, one for each server.
  
  
• You create an SNMP group for Interface monitoring with four SNMP nodes, two in/two out for each server.
  
  
• You use the Anchoring extension, select the "Ping A" node and tie all the SNMP ones to it.
  
  
• You use the Anchoring extension, select the "Ping B" node and tie all the SNMP ones to it.

Once you reload the dashboard, you'll be able to click on the "Ping A" and "Ping B" nodes to toggle their related nodes' visibility. That's what this extension was made for :)

Still confused?

If you're questioning why there are no screen shots here... well, please be patient. I'll add them later on.

If you need help, don't hesitate on posting a support request here at the forum.