What you are looking for is alert correlation.  SCOM does not have alert 
correlation across multiple health services.  You'd have to architect a 
solution that handles the criteria.  Ideas:


1.       Orchestrator workflow, look for low level informational alerts or 
monitor states, then have the runbook remediate, check for other server 
services, etc... then raise an alert if necessary.

2.       Custom MP, agentless WMI monitors from a watcher node, feed them a 
datasource of servername and service name, have them handle the correlation in 
their monitoring datasource (script).

3.       Command channel notification subscription to low level info alert for 
a service, that inspects the other server(s) via remote WMI and then modifies 
alert to critical, adding context, if necessary.



From: [email protected] [mailto:[email protected]] On 
Behalf Of Pete Hakesley
Sent: Monday, October 14, 2013 8:45 AM
To: 'msmom ([email protected])'
Subject: [msmom] Question about Windows Service Monitoring

Hi all,

We had had a request to Monitor a Windows Service on two nodes  - no problem 
here 'bread and butter stuff'

However they operations group only want an alert if both nodes have the service 
failed but no alerts if only one from two fails.

Other than using a DAD model and suppressing the alerts at the component level 
which would work but to do this for some 20+ Windows Process is there a better 
way in SCOM to achieve this?

Thanks in advance.


Peter Hakesley PJH2711-RIPE, CDCT, MBCS

SCC Services

Data Centre Operations - Service Implementation - Enterprise Systems Engineer

SCC Cole Valley 1,
20 Westwood Avenue,
Tyseley,
BIRMINGHAM B11 3RZ

Tel: +44 (0)845 351 0680 xtn 4006
eMail: [email protected]<mailto:[email protected]>
Web : www.scc.com<http://www.scc.com/>

[Description: Description: BCS_Logo]




<<inline: image001.jpg>>

Reply via email to