Craig McMurtry, Longhorn Server App Platform Evangelist, posted a great list of ways to manage/monitor WCF services to an internal email discussion last week. It was so good I asked Craig if I could post it for both of my faithful readers to read. Below is a great list and if this doesn't make you get even more interested in WCF, nothing will. Thanks again to Craig!
Eleven Ways to Manage/Monitor WCF
First, there is a Configuration Editor, and there has been since the PDC â€™05 build, so no, you donâ€™t have to â€œedit cryptic config files with Notepad.â€
Second, there is a rich set of performance counters for measuring every aspect of a generic servicesâ€™ performance at the level of the service, its individual endpoints, or its individual operations. And you can readily add performance counters for monitoring what is not generic.
Third, there is a WMI provider with a rich set of classes in the root\ServiceModel namespace allowing one to discover, interrogate, monitor and script running services using whatever WMI-enabled tools suit the administrator, from Tivoli and OpenView to ScriptOMatic and PowerShell.
Fourth, you can log messages at any or all of the stages in the reception and transmission process.
Fifth, you can activate activity tracing either at deployment time, or, through the WMI provider, at runtime. The volume of activity tracing can be set to various levels between â€œcritical exception notificationsâ€ and â€œverbose,â€ and since the product team uses the activity traces for debugging themselves, one can be confident that the activity traces would allow one to diagnose defects without having to fire up a debugger.
Sixth, there is the magnificent Trace Viewer, which not only serves as a viewer for message logs and traces, but also allows one to see, graphically, the thread of execution as it progresses from node to node across the network.
Seventh, you can sit on top of the activity tracing system and add custom tools like the Live Service Trace Viewer, pictured below, which shows interaction among endpoints at runtime.
Eighth, security events are audited, with a configurable level of detail, not only into the activity log, but also into the event viewer.
Ninth, in the process of defining business endpoints, designers and architects can readily add endpoints dedicated to exposing information pertinent to administrators.
Tenth, you can use the free Management Model Designer to assemble a comprehensive management/health model for your administrators from the preceding nine facilities.
Eleventh, you can incorporate your management/health model into an integrated management environment like Tivoli, OpenView or MOM.