Select Page

In my KSCOPE12 presentation,  Planning for and Managing Hyperion Infrastructure, I give a 100,000 foot view of monitoring.

In this article I bring monitoring down to a 20,000 foot view.

There are many products on the market to choose from for monitoring your Hyperion system(s).

In a nut shell you need a solution which has the below capabilities and characteristics:

  • Little or no false alarms
  • Ability to monitor: Database, Disk, Processor, Memory, and Network
  • Ability to extend base product capabilities with custom monitoring
  • Ability to test key parts of the system to prove it really is working

I narrow the field down to five that I believe have one or more of the following qualities: easy to implement, inexpensive, and/or purpose built for Hyperion.   You may note I have excluded solutions from: BMC, IBM, HP, and CA which have products not exhibiting the former qualities.

When implementing any new software you should go through an evaluation cycle to understand how one tool compares to another in terms of capabilities and price.

The Problem

You understand Perl scripting and want to deploy a quick solution to monitor which is free and has no real infrastructure needs and can be deployed in less than an hour.

The Solution

The below Perl code was a weekend project several months back.  It does have a feature (testing of Smart View logins)  you may wish to integrate with a more formal monitoring solution.

It is easy to implement and can run from your existing Hyperion deployment with no other software.  It relies on Perl and will work with the versions available on the 11.1.2 Hyperion line.

Modify the configuration file to suite your environment.   This means changing the server names to correspond to those which are applicable to your environment, adding or removing services, adding or removing log and keywords, setting your system outage times, and finally scheduling the script to run on an interval (usually every 5 minutes).

Sample config file:

Perl script:

John A. Booth