Tuning Alert Noise from Exchange SLA Mgmt Pack Rules

I ran into a problem with a rule in the Exchange SLA Scorecard that is both generating alert noise and possibly impacting the availability numbers of the scorecard unnecessary. Other folks saw this issue as well, and posted the question in the MS newsgroups. I provided an answer to reduce the noise, but now would like to offer up (and justify) a better workaround. Your comments and feedback would be appreciated. The impact comes from the Exchange SLA MP Rules for Exchange 2003\Service Control Manager Start\Stopped rule, which fires alerts based on Event ID 7036, which is the event generated when a Windows services stops normally. The Problem... This is problematic because the WinHTTP Proxy and WMI Performance Adapter services are set to startup type of MANUAL by default and thus stop and start normally during normal operation without impact to messaging service, and set this rule off repeatedly (and quite unnecessarily)…..ultimately negatively impacting bottom-line front-end availability. The Original Workaround… Initially as a fix, I added the following regular expression (see below), which causes this rule to only fire when the service stops unexpectedly Event Number | matches regular expression |^(7021|7024|7031|7032|7034)$ .....logical in most any other circumstance, but in this case, it appears it will impact the performance numbers of the Exchange SLA Scorecard because front-end server availability is calculated based on occurrence of the following events (desription below taken from the Exchange SLA User Guide). Exchange Front-End Server Availability | Exchange Server 2003 | Events 7036, 6006, 6005.    A Better Workaround… I never want to post a fix that impacts the accuracy of the information being reported, so I wanted to offer what is my mind a much easier and more appropriate fix in this situation. I mentioned the WinHTTP Proxy and WMI Performance Adapter services are set to startup type of MANUAL by default and thus stop and start normally during normal operation, so simply set the startup type on these two services to AUTOMATIC. This squelches the alert noise AND removes any impact to the accuracy of Exchange SLA reporting. The original newsgroup thread where we discussed this topic can be found here. Powered by Qumana