Going forward, are there tools that integrate Health Monitoring (today, circa .NET 4.5) with things like System Center Operations Manager? It''d be great if we could somehow use the SCOM UI to aggregate ASP.NET Health as defined by Health Monitoring.

FWIW - I think you did a nice job, and it was very impressive when it came out back in 2005.

Thanks for checking in. Interesting to hear about using HM for visit logging. I would also suggest you check out some of the new web analytics providers (E.g GA, Chartbeat, Mixpanel, etc) that provide huge analysis value add on top of anything that is reasonable to build yourself anymore.

To answer your questions:

(a) Does LeanSentry use HM?
LeanSentry gets its data from a variety of sources, one being HM but also performance counters, WMI, IIS logfiles, etc. Its just not possible to get everything in one place if you want to create a hollistic picture about your apps''s health and performance.

(b) Do we use HM events?
We use ASP.NET health events as part of our analysis as needed. Right now, we mostly focus on errors.

(c) What is the difference between HM and LeanSentry?
I would say LeanSentry is a completely different tool. HM is a decent eventing framework you can use for your own purposes. LeanSentry is meant as a comprehensive monitoring and diagnostics tool that abstracts the health monitoring data sources, and focuses on delivering the critical insights to help improve application health.

I dont want to write a whole lot more here, but definitely check out our How it works page: https://www.leansentry.com/Howitworks.

Best,
Mike

a) Where is this information gathered from? Is/would it be coming from HM events/error descriptions (I''ve added custom description information to the default ''unhandled'' error)?

b) You mentioned above that some information could ''only'' be retrieved from HM...I think I had read that earlier, which was another reason we worked to get HM plugged into our system instead of a home baked or third party logging system...how is that information represented in Lean Sentry?

c) Finally, based on your answer for a) I guess, are you meant to use HM and Lean Sentry in tandem or is LS supposed to replace HM?

Thanks in advance.

In II6 (and IIS7 classic pipeline), ASP.NET creates a WindowsIdentity to represent the authenticated user. That identity is never disposed, so when the buffered event is formatted, the identity is still valid.

In IIS7 integrated pipeline, the WindowsIndentity is created by IIS and passed to ASP.NET, and it''s disposed when the request is completed. When the buffered event is formatted, the identity is no longer valid, and you get a "Safe handle has been closed" error.

The only solution seems to be to turn off event buffering.
http://forums.iis.net/t/1172849.aspx

Going to take a look at Lean Sentry, thanks for sharing that.

Re: WCF, I agree. To be honest, we have used it internally a number of times and ended up regretting it most of the time. If you have the time to spend, we found building a more low level communication solution (e.g. REST over HTTP with HttpWebRequest) is almost always more reliable, and in the end actually faster because you dont spend any time working around any WCF weirdness. WCF is a classic example of overengineering that didnt focus on delivering basic developer value.

You know what else is a great example of MSFT overengineering? WWF. Dont get me started.