We use it in combination with Solarwinds Orion and the Enterprise Console which for the price is very feature rich and well supported. I will agree that products such as HP Openview and Ciscoworks are excellent but only realy come into there element when monitoring there own vendor specific equipment.
With regard to OpenView;
The problem with OpenView is that an out of the box implementation will do sweet FA. OpenView will manage devices from any vendor in 99% of cases.
Additionally. OpenView is a brand name for a wide range of HP Enterprise Software products (and following the mercury acquisition all HP enterprise apps must drop the OpenView prefix) and does not refer to one particular piece of Software.
Common pieces of OpenView Enterprise Management software are:
- HP OpenView Network Node Manager: In its latest release is now known as NNMi
- HP OpenView Operations Manager for Windows\Unix - Now known as OMU (Ops Manager for Unix) and OMW (Ops Manager for Windows)
Subcomponents of these are performance management tools, reporting tools, messaging tools, and smart plug-ins to manage everything from ESX to rickety old Tru64 Boxes (nearly all stuff is catered for).
These enterprise management products are simply a framework to enable you to build your own monitoring solution. They require reading of a phenominal amount of concept guides and whitepapers but can pay off incredibly well. I administer an Operations Manager for Unix install, it provides service dashboards, a trouble ticket interface into multiple service desks, report generation, sla monitoring and server management/monitoring. The application allows you to install "agents" on almost any OS platform (as well as snmp for stuff like SANS, Aircon units, Tape Units, UPS's etc), these agents then talk back to the central management server via HTTPs. Through these agents you can then execute scripts and applications, browse things such as WMI, look and test for almost any condition and read any logfiles etc. This information is then turned into an alert, or stored in the DB as a performance metric, or promoted into a trouble ticket. You can also run automatic commands based on the conditions you find, such as restarting a service, deleting the temporary files, kicking idle users off terminal service etc.
Its a very powerful system and can massively reduce the need for a big Ops team when managing large estates. Quite expensive though, agent licences are tiered to sockets, OMU requires Oracle (most products can run on x86 windows iron with whatever DB you like), for example the install I administer requires two HP SuperDomes to manage round 2000 servers and a thousand SNMP devices. An additional 4-Way SQL box is used for Reporting.
A lot of companies can find Network Node Manager or Operations Manager on a shelf somewhere - even if its 10 years old, reactivation of support will get you the latest version - though any increase in licence capacity would obviously have to be factored in. It does require a large amount of effort to get it to go beyond the basics, but worth it if you've gone to the expense of buying it!
