Best Practice for Alerts/SNMP Traps

[kutner]

I'm looking for a document or help deciding what events should be alerted on.  In general I think any alert of 'error' or higher would be something that I want to make sure I get notified about, but I don't know what will generate an 'error' or higher.

Similar question for SNMP trap generation and monitoring.    I have a customer that wants to know how to best utilize a 3rd party site-wide monitoring tool.  They are trying to decide what should be monitored via Ops Mgr and whether to pass along snmp traps or have traps from the FAS go direct to the site-wide monitor.

Are there any documents or guidance available?

Thanks,

Andy

[kravi]

I think if you are looking for general guideline, you should be look for

warning or worse events. Otherwise, it depends on what you are looking

for, how you have configured various thresholds. Usually, there are two

types of thresholds in Operations Manager e.g. Volume Full Threshold and

Volume Nearly Full Threshold. When "Nearly" thresholds are breached,

warning events are generated and when next threshold is breached error

events are generated.

You can make Operations Manager send traps for all events to your

3rd-party site monitoring tool as well. So, it is upto you which tool

you prefer.

I don't think there is any guideline as such for any of these things.

[fenton]

I have seen some customers want to use OpsMgr as a trap consolidator so that they have control of what central operations see and alert against - in these environments all FAS based traps will be sent to DFM and then forwarded based on severity or the type of controller;  The major factor for them to take this approach was:

a, to retain control

b, stop the monitoring framework from getting swamped with lots of spurious traps

I have had other customers that also send FAS alerts directly from FAS as they don't want DFM beign a single poitn of failure for alerting...so I think it's very much a horses for courses type scenario;