On Win2K, on the Services plugin you can bring up properties of each service. There's a Recovery tab. It allows you to set up what will happen after failures.
Please define precisely what "failure" means in this context. Does it mean just that the service threw an exception, or does the OS do some kind of active monitoring of particular performance characteristics of the service? What are all the different things that can cause a "failure"?
Is there anything I need to do in my Windows Service code to enable or support those recovery features?
I'd love to get some links to documents (Microsoft and other) that go into this in detail. I'd also like some detailed personal feedback from someone who has direct experience in this area.
( For those whoe aren't familiar you can find the Services plugin like this: Start / Settings / Control Panel / Administrative Tools / Services. )
Context: I'm currently writing a couple Windows Services using Visual Studio.NET 2003 (.NET Framework 1.1). I'd like to make them reliable, and also configure the OS to monitor them for additional reliability.