Develop Performance Interface.

From a global standpoint, performance analysts and DBAs benefit from having the following groups of information at their fingertips:

System-Level

Contention information – where the server has been bottlenecked and waiting to accomplish end user or system requests. Includes waits, locks, and deadlock activity.
SQL Execution metrics – includes summarized or detailed capture and reporting of issued SQL statements and their performance metrics (scans, sorts, etc.)
Time model – where the server has been spending its time; what activities have consumed most of the engine’s attention.
I/O statistics – metrics concerning all I/O activity at the system, database, object, and issued-SQL levels
Resource usage – metrics concerning elements such as memory usage, CPU, and space consumption
General performance metrics – everything from network activity, failed logon attempts, and other such things.
Configuration info – how the server has been set up and the configuration settings of its defined parameters
Operational data – statistics such as system start/stop times, encountered errors, etc. (many times comes from reading a database’s error/alert log)

Session-Level

Contention information – where user connections have been bottlenecked and waiting to accomplish their requests. Includes waits, locks, and deadlock activity.
SQL Execution metrics – includes per-connection capture and reporting of issued SQL statements and their performance metrics (scans, sorts, etc.)
Time model – where user connections have been spending their time; what activities have consumed most of the each connection’s attention.
I/O statistics – metrics concerning all I/O activity at the user connection level
Resource usage – metrics concerning elements such as memory, CPU, etc., for user connections (if available)
General performance metrics – everything from network activity, etc., for each user connection.