I misremembered. I thought there was a network monitoring solution in the huge 12.0 appliance library. I'm sorry if I'm missing it. I've been looking for something to monitor and management the ridiculous number of servers running in our laboratory. Some are mad science; others are kind of vital.

I came across observium, and it looked perfect for our needs, so I thought I'd take it for a spin with TKL LAMP.

Observium is an autodiscovering PHP/MySQL/SNMP based network monitoring which includes support for a wide range of network hardware and operating systems including Cisco, Linux, FreeBSD, Juniper, Brocade, Foundry, HP and many more.

It's distributed via svn; installation is a breeze (there's an additionaly note for squeeze). I'm imagining a patch solution for it, and it's not really overtaxing me. It's a standard lamp patch. The cypto-inithooks are no longer so crypto. So a turnkey solution for the web app is just a matter of being able to create and admin user in the sql file and then setting a proper password for it on first boot.

The installation instructions are here for Ubuntu and Debian5/6 are here: http://www.observium.org/wiki/Ubuntu_SVN_Installation. In all honesty the install directions are so good I don't suppose there's a need for build notes.

Anyone have experience with observium?

Forum: 
Eric (tssgery)'s picture

I have no experience with observium, but it looks pretty nice.

 

I've been looking for something like this and have been thinking about Nagios or Zenoss but would be happy to try out your patch.

Eric (tssgery)'s picture

I wanted to try this out so I built a patch for it. I opted to build it on top of the core appliance as opposed to the lamp appliance. I've installed it myself and added a few of my servers to it in order to try it out. Feel free to try it out and let me know how it works.

The source is in github: https://github.com/tssgery/observium

The patch and iso are on my personal websile:

patch: http://www.aceshome.com/observium.tar.gz

iso: http://www.aceshome.com/turnkey-observium-12.0-squeeze-x86.iso

Jeremy Davis's picture

Good on you guys. Observium looks pretty cool. I'll take you patch for a spin when I get a chance Eric.

I look forward to trying it in our lab.

Fantastic. Glad to see that you picked this one up. Patched effortlessly. Less effortless is configuring the servers to be observed. I'd love to see a patch to bring servers to a state ready to be observed. I've scripted part of it. But the application monitors are very tedious (to me) to configure.

Eric (tssgery)'s picture

If you can share the script you've started to configure the servers? Maybe I can take it and extend it to provide something more helpful.

Here's a script intended for a Raspberry Pi, but I'm pretty sure it's of use to Debian and Ubuntu targets. It's on github: https://github.com/ghoulmann/observium-raspi-monitoring

Eric (tssgery)'s picture

I've performed the same steps on several machines myself :)

 

What I was thinking was to add these to the observium patch and add a wrapper script around the observium 'addhost.php' script that remote shelled into a server, checked to see if snmpd was already installed. If snmpd was not installed then to run these steps before calling addhost.php.

The other option is to make people ssh into the servers themselves and pull the package from the observium server and run it manually.

Of course, I'd need to add support for non-debian based distributions as well but that should be pretty trivial.

To the observium host machine makes sh scripts available? Perhaps we could wget the scripts from that instance and pipe them through sh on the new client?

Eric (tssgery)'s picture

I see to usage patterns:

1) While logged onto the oberservium server, run:

        addHost hostname/ip password

This would ssh to the hostname, configure it remotely, and add it to the observium inventory

 

2) While logged onto the device to monitor, run:

        wget -O - http://myhost/setupdevice | /bin/bash

Then you would need to return to observium and add the device to monitor

 

The good news is that the same script can fulfill both uses. 

I have some ideas for this but it wont happen really quickly, if someone else beats me to it I will not complain.

 

Hi all,
 
Can any one please confirm about the mentioned below detials are supported
over observium .
 
 
   - Interface monitoring ( All ypes of pors as STM's, FE , GE , XGE , etc
   - device statisics  (CPU, Memory , etc )
   -  IPSLA MIB support
   - QOS/COS support
   - RRD suuport
   - Format of files relaed to data of graphs
   - VPN/neighbour  prefix details
   - Status of Routing protocols
 
 
-How much of Hardware requirement will be needed for monitoring of about 50
Routers along with  150 L3/L2 switches considering polling time for about 5
min. As an average routers have about 20 ports nd L3 switches with about 40
ports eachs, nd layer switches have 25 ports.
 
 
  - How actually network discovery works for observium ? I mean do I have
   to just enable SNMP with same community and same Snmp version  on all my
   devices so that observium can discover my whole network and does observium
   also show me the whole map for my network ( connectivity b/w my devices)
   and most important do i need to make any device as Seed device for my
   network ? from which it will automatically penetrate to whole
   network...Further more after discovering of my device can i filter those
   graphs so that I can only plot useful or main graphs
 
 
  -  Can I also make groups to those graphs I mean is it possible that I
   can make a separate tab for my roters and the \n separate tab for switches
   etc.
 
Thanks
 
Muhammad Rizwan Azher

Add new comment