The System Administrator has it easy compared to the Application Administrator, the former have statistics to gather, services to test, thresholds to watch for, and SLAs to meet but the latter works with perception and with processes that are difficult to measure. Where systems are designed to be as data independent as possible, applications are data dependent almost by definition. One "document" may not the same as another, even if they may display identically to a user, let alone when they are talking about completely disparate subjects. Application / Solution Health Monitoring encompasses:

• Standard availability; can the intended end users of the application access it when they need to
• Authentication; does whatever authentication service associated with the solution function
• Authorization; does whatever authorization service associated with the solution function
• Sub-component monitoring' third party application and services as well as sub-processes of the whole solution. For example:
  
  • A database connection and authorization
  • A search service and an index service
  • A notification service
  • An administrative service
  • Email access (send and receive)
• Various ways to provide input to the system (ie: email vs web vs desktop client)
• Various workflows, dataflows, and other processes involved in the application

Application / Solution Performance Monitoring encompasses both the notion of server-side performance and end-user performance. The former is more like typical system's level monitoring, it is usually very statistically oriented and mathematically calculated. So too is the end-user performance but because there are so many variables outside of the application's control (like network bandwidth and latency, client configuration, and data dependency) the monitoring has to be broken out to show the parts that are and are not within the sphere of influence for the application.