Page Life Expectancy (PLE) is a metric often used to determine the health of the SQL Server buffer cache. A low PLE value indicates pages are not staying in cache very long and could indicate a variety of issues. My experience is that many people alert on a low PLE value, but then do not have the data to understand what caused the problem. Also in my experience, PLE is often affected by a query or set of queries that read a large amount of data from disk that in turn age out other pages from memory. Would it help if DPA collected information about the contents (tables/indexes) of the buffer cache periodically (and stored historically) and display that along with the correlating SQL statements. In other words, provide a better indication of which query caused the PLE to suffer.
