Wednesday, November 20, 2013

Nagios in an Enterprise Environment

If you are looking for a network monitoring solution using Nagios or Nagios XI to monitor a very large number of hosts, I am talking more than 5000 hosts, it is possible to design them using Nagios or Nagios XI (hereinafter known as Nagios).

There are several critical points you need to take into consideration. Two very important points, this will either make your design flawless in terms of performance or give you a headache no medicine can cure!

Point 1: Hardware specification - Take into consideration CPU, RAM, IO of HDD, the higher the better (SSD even better)

Point 2: Frequency of check, how data is handled (queuing and writing)

The solution we designed writes over 1,000,000 records per day on a MySQL database (yes the DB needs to be tweaked for performance! and regularly checked for potential issues).

We designed a complete new dashboard from ground up. This dashboard displays only alerts, has the following action points:

  1. Able to assign alerts to agents
  2. Able to assign alerts to level 1, 2 & 3 engineers
  3. Assign ticket ID to alerts
  4. View hosts related information (asset information; like location, purchased date, warranty information, support engineers to call, etc)
The main interface was also design from ground up. This interface allows users to add hosts and select pre-defined templates for that particular hosts. We created over 100 templates, ranging from a switch, routers, firewall, load balancer's, multiple OS's, oracle, DB2, *SQL,and many more.

This is the high-level design that catered for this requirement.

Have a good week ahead!