IBM is currently launching web service management tools into the WebSphere/Tivoli space, including a web services dashboard. This now appears as two slightly different products: IBM Web Services Navigator and IBM Tivoli Monitoring for Web Services.
The web services dashboard displays the interaction patterns of web services, and these can be tracked against predefined profiles. A service profile is a recognised pattern of service interactions (an orchestration scenario). A service profile distribution is a statistical pattern indicating the occurrence of service profiles. (For example, this might allow the system to generate an alert if a given exception condition is executed for more than x% of cases. More generally, it supports Statistical Process Control.) Deviation from accepted service profiles or profile distributions may generate an alert, demanding someone's attention. Alternatively, the alert may be passed as a signal to the web service platform, where it triggers some autonomic response.
A major issue for any tool of this kind is scaleability. A screenful of information may be good enough for a small demonstration, but how does the tool cope with larger and more complex situations, when the information would spill over many screens?
This is precisely why you need an architect, who can break the information down into manageable chunks. In a separate post on Web Service Management, I discuss the role of the service architect, and the web services management process.