Database "exec avg" time inconsistencies..

As i'm getting such excellent support from this forum, i'd like to discuss / brainstorm another "weird" scenario in my Dynatrace-instrumented app.

I'm using DT to monitor a .Net ETL job. To simply, it does the following:

1) Call a SQL 2012 sproc, that returns 150K rows of data

2) Process above data

3) Write out summarized data to destination.

I'm having difficulty correlating the "database exec avg" value for this sproc in the "Database" dashlet.

I'm using ADO.NET / .Net 4.0.

I'm also using the .Net Enterprise Library for calling the SQL sproc.

So the code kind of looks like:

CallSproc()

{

SET sproc params into DBCommand

public virtual System.Data.DataSet ExecuteDataSet(System.Data.Common.DbCommand command)
    Member of Microsoft.Practices.EnterpriseLibrary.Data.Database
CALL public virtual System.Data.DataSet Microsoft.Practices.EnterpriseLibrary.Data.Database.ExecuteDataSet(System.Data.Common.DbCommand command)

ie : [http://msdn.microsoft.com/en-us/library/bb748692(v=pandp.50).aspx|http://msdn.microsoft.com/en-us/library/bb748692(v=pandp.50).aspx]

Source code: http://entlib.codeplex.com/SourceControl/changeset/view/99885#527021

Above method is explicitly instrumented in Dynatrace. The purepath shows the following (partial output):

So ExecuteDataSet took almost 15 seconds to complete. This is consistent with SQL Server timings, via say SQL Profiler Trace, or manual execution of the sproc from Query Analyzer.

But, the "Database Dashlet" says something else completely! If i drill down to "Database" from the above purepath, i get:

400 ms "exec avg" does not match 15 seconds of execution time.

So what exactly is being measured/reported by the "Database" dashlet?