Average # monitored Components?

Enginseer

Hola,

We're a long time Orion customer, but first time APM customer, and I'm wondering what's considered "average" for application monitoring for most people?  We've got the ALX package and monitoring abour 350 windows boxes.  First thing we did was set up a general template monitoring about 8 services that run on all Windows 2003 servers and applied it to our 350 hosts...

That's a shade under 3000 monitored elements, and it took our server which previously monitored 2500 NPM objects to it's knees performance wise...

So my question is... what do people normally monitor?  Do you monitor "everything", i.e. general health? or just really specific things?  I just want to make sure my idea of what I want matches what's actually possible and done in the real world...

 

Thanks,

Brad

qle

I imagine this would be unique to each environment. For example, there may nodes that you simply monitor for up/down status. Then there are other critical systems where you need to know the moment the application is unavailable even if the system stays up.

I'm assuming that the 8-service template you used was the Windows Server 2003-2008 application template that is included with the current version of APM. My opinion for the best use of this template would be for the following:

• Reactive: systems you believe are underperforming but you need hard evidence to back up this theory and find root cause
• Proactive: systems whose performance are critical to your entire organization's operation such as ERP

In the end, if you decide this applies to all 350 systems then it looks like you'll need to budget for a beefier system to handle NPM.