Open Source Network Monitoring Tool

[wahmed]

After lot of test, trial I went with Zabbix. I configured Zabbix to add quite a few local and remote networks/devices and it has been running for last few weeks without issue. Things that takes forever in Icinga(due to complexity and my short knowledge) I was able to get those going in Zabbix in no time. I even created custom made template for Netgear 48 port smartswitch and now can monitor every single port of it with both graphs and numbers all from Zabbix. I even get Jabber text msg from zabbix to my android when there is an issue with anything.

May be i have not came across anything yet where Icinga/nagios shines over Zabbix greatly, but what i needed to do to monitor networks spread across 2 regions with multiple locations, Zabbix had been amazing. I cannot thank enough everybody for your input on what you guys use which allowed me to see what options were there.

 

[schannall]

I use OMD and like it quiet a lok.

sC

 

[raj]

After lot of test, trial I went with Zabbix. I configured Zabbix to add quite a few local and remote networks/devices and it has been running for last few weeks without issue. Things that takes forever in Icinga(due to complexity and my short knowledge) I was able to get those going in Zabbix in no time. I even created custom made template for Netgear 48 port smartswitch and now can monitor every single port of it with both graphs and numbers all from Zabbix. I even get Jabber text msg from zabbix to my android when there is an issue with anything.

May be i have not came across anything yet where Icinga/nagios shines over Zabbix greatly, but what i needed to do to monitor networks spread across 2 regions with multiple locations, Zabbix had been amazing. I cannot thank enough everybody for your input on what you guys use which allowed me to see what options were there.

Hiya

Would you share how you configure zabbix pls

 

[wahmed]

Would you share how you configure zabbix pls

The configuration is nothing special really. Following the installation instruction for Debian/Ubuntu in this link:
https://www.zabbix.com/documentation/2.2/manual/installation/install_from_packages

Installed Zabbix Agent on all Proxmox+Ceph nodes, added them as host into Zabbix. From there its basically fine tuning such as i wanted Zabbix notification whenever CPU consumption went up over 70% and disk storage space fell below 30%. I installed it on a node on its own outside the main network in a different location. So that i get notification even when Internet connection fails in the site. I went with Zabbix 2.2 LTS so i dont have to update often.
I also configured Jabber so that i get SMS notification in my android. Only thing i am trying to configure is to monitor Ceph MONs and OSDs individually.

 

[c0urier]

I'd say go with LibreNMS it's a fork of Observium with a build in alarm system and support for Pushover.

 

[Phinitris]

Hey,
if your switch supports sFlow, Netflow I recommend FastNetMon. It's an DDoS analyzer and can store metrics in Graphite.

https://github.com/FastVPSEestiOu/fastnetmon
Example of Grafana: https://cloudrouter.org/cloudrouter/2015/07/09/fastnetmon.html

Best regards
Henry

 

[fireon]

We use Nagios since ... a long time. And it works fine! For configuration we use Lilac-Reloaded: https://sourceforge.net/projects/lilac--reloaded/ It is an template based webconfigurator. After you added your own templates, it is easy with some clicks to add your proxmox host.

2 Examples are attached.

A second one what is really importend and simple to configuration ist Smokeping: http://oss.oetiker.ch/smokeping

 

[Ofca]

Can't believe nobody mentioned netxms - most thought-through monitoring software, once you understand how it works

 

[iconvergence]

+1 observium is very good

 

[wahmed]

Really awesome info everybody! I got to know some of the network monitoring tool i never heard of. I already tried some of it. Currently i am comparing all other tools with Zabbix. For me Zabbix some what became the benchmark standard. I am beginning to compare all other tools to it. I configured quite a few devices in Zabbix. So far it is doing a great job monitoring.
I did like the polished great GUI of Observium. But did not like the fact that i cannot add a device purely with IP address. It is very important. Really good to see there are many options for monitoring and actual users opinion. I will definitely be keeping an eye on this thread.

 

[maca87]

Hey Wasim, it would be great if you share the result of your comparison! I am still using Pandora FMS and I would like to know how things are going
Thanks

 

[wahmed]

it would be great if you share the result of your comparison! I am still using Pandora FMS and I would like to know how things are going

I am still using Zabbix and loving it! Over last few weeks i got to know more about Zabbix and added numerous networks to it. From monitoring to graphing to notification, all working just great. Recently i added Hard Drive temperature monitor for ALL drives with max temperature threshold. Now i can monitor the temperature ups and downs over long period of time down to each drive. All simply using Zabbix and smartmontools.
Attached is a snippet of one Ceph node with 8 drives and their temperature graph over 14 days period.
After using Zabbix for so many happy days my opinion may be biased but features that i lacked in other monitoring tools that i tested, Zabbix fulfills them all. Currently with one Zabbix i am monitoring about 16,000 items for 116 nodes/switches/equipment. About twice more are still pending to be added. Also the ability to create Live network diagram in Zabbix which shows online/offline/issue of equipment even down to the network connectivity is simply great.

I was new to Nagios/Icinga/Zabbix at the time i was looking for good monitoring tools. What i accomplished in icinga after trying it for 5 hours straight, i accomplished the same plus more in less than 45 mins with Zabbix. Thats just from my personal experience. I know Nagios/Icinga and likes are very powerful. But Zabbix has fulfilled the need for monitoring our large network deployments just fine thus far.
Hope this helps.

 

[Q-wulf]

seeing how you run ceph, you probably wanna look into graphing Smartmon/smartctl values.
Afaik Zabbix 2.2 can do that via a plugin.
Very easy to spot failing OSD's by their trends. So you have something to do during your maintenance windows

 

[wahmed]

seeing how you run ceph, you probably wanna look into graphing Smartmon/smartctl values.
Afaik Zabbix 2.2 can do that via a plugin.
Very easy to spot failing OSD's by their trends. So you have something to do during your maintenance windows

All OSD graphs are created from smartmon values including the temp. graph i posted earlier. I am pulling varieties of data such as read/write/sector errors etc. into graphs. Is it what you are talking about? So far it helped us identify 6 aging OSDs which probably would have gone unnoticed without the ability to set threshold notification.

To add to my previous post, the graphing ability also helped me to play with temperature control as you can see from the graph. We cool big part of the datacenter using outside freezing air. The big temperature drops were result of pushing more cool air in.

 

[Q-wulf]

I am pulling varieties of data such as read/write/sector errors etc. into graphs. Is it what you are talking about?

Yes, thats what i mean, its the best thing ever to monitor 5.2k Disks in total on roughly 94 nodes, that way. Gives the minions something to do

 

[ShaneCar]

I noticed many mentioned Zabbix or Nagios. Zabbix and Nagios are very reliable, time-tested, and quite robust for open source tools. Zabbix might give you more visualization through time-based graphs, however, Nagios will offer more plugins for versatility. You might also find more online for Nagios than Zabbix. With either, I would look into adding on an alert correlation config/ tool. Both of these monitors can be very high maintenance if you set up email or sms notification to keep you aware of your network health. BigPanda, is a popular correlation tool, which for example, will sit on top of Nagios and seamlessly combine all related alerts into single incidents- saves time and brain power.

 

[RobFantini]

I am still using Zabbix and loving it! Over last few weeks i got to know more about Zabbix and added numerous networks to it. From monitoring to graphing to notification, all working just great. Recently i added Hard Drive temperature monitor for ALL drives with max temperature threshold. Now i can monitor the temperature ups and downs over long period of time down to each drive. All simply using Zabbix and smartmontools.
Attached is a snippet of one Ceph node with 8 drives and their temperature graph over 14 days period.
After using Zabbix for so many happy days my opinion may be biased but features that i lacked in other monitoring tools that i tested, Zabbix fulfills them all. Currently with one Zabbix i am monitoring about 16,000 items for 116 nodes/switches/equipment. About twice more are still pending to be added. Also the ability to create Live network diagram in Zabbix which shows online/offline/issue of equipment even down to the network connectivity is simply great.

I was new to Nagios/Icinga/Zabbix at the time i was looking for good monitoring tools. What i accomplished in icinga after trying it for 5 hours straight, i accomplished the same plus more in less than 45 mins with Zabbix. Thats just from my personal experience. I know Nagios/Icinga and likes are very powerful. But Zabbix has fulfilled the need for monitoring our large network deployments just fine thus far.
Hope this helps.

Hello symmcom
I'm trying to get zabbix configured and looking for some advice .

Which version are you using? I'm trying 2.4 from http://repo.zabbix.com/zabbix/2.4/debian/ jessie , however the documents at https://www.zabbix.com/documentation/2.4/manual do not seem to match my version.

Any install / setup advice would be appreciated.

 

[Q-wulf]

Which version are you using?

went with Zabbix 2.2 LTS so i dont have to update often.

I'm using Zabbix 2.4 here at home. 2.2 at work.

advice:
- Do not use slow VM storage
- get the Proxmox plugins
- get a smartctl/smartmon plugin and monitor your Cluster wide smart capable drives (saves you digging after dark when a drive fails)
- get as much data about your network connections as you can when given the option.

 

[RobFantini]

for the server, did you use debian packages, or ones from zabbix ?

 

[Q-wulf]

for the server, did you use debian packages, or ones from zabbix ?

Zabbix package from their download page for zabbix 2.4 on Ubuntu Version 14.04 LTS (Trusty).
Followed the general Install instructions here:
https://www.zabbix.com/documentation/2.4/manual/installation/install_from_packages

From there i followed along with a combination of their documentation, their forum and google.