pegasus statistics right now works on a submit directory, where it picks up the workflow uuid from the braindump file in the submit file.

It would be nice to extend it to take multiple workflow id's as input . Is useful to get statistics across multiple root workflows that are stored in the same database.

Input
------------------------------
pegasus-statistics should take in the database and optionally the list of root workflow id's for which to generate the statistics

Statistics Generated
-------------------------------
We will generate statistics as

Job breakdown ( No of jobs succeeded / failed etc )
Sub Workflows ( No of sub workflows run, succeeded/failed etc )
Task Breakdown ( need to have a switch for it ? as can take a long time)

Workflow cumulative job wall time
Cumulative job walltime as seen from submit side :

Job Breakdown statistics by transformation type ( similar to the jobs breakdown file )

Time Statistics
Number of Jobs/Invocations run per month/day/hr