Active Session History (ASH) is a time-based sample of session activity in an Oracle database. It is a powerful source of information on how time is spent in the database, however misconceptions about and misuse of ASH data are common. Properly using ASH to understand and diagnose database performance issues requires understanding both its architecture and the theory of time-based sampling. This two-part presentation will give an overview of ASH architecture and the theory and methods for using ASH:

Part 1
• The ASH mechanism: sampling, defaults, controls
• What an ASH row represents and the multi-dimensional nature of ASH data
• Estimating DB Time from ASH using SQL
• Top Activity and ASH Analytics interfaces in Enterprise Manager

Part 2
• The “fix-up” mechanism for TIME_WAITED and other important values
• Estimating event counts and average latencies from ASH
• Finding outlier events in ASH and the risk of sample errors
• Visualizing DB Time and event latencies