Name

pegasus-statistics — A tool to generate statistics about the workflow run.

Synopsis

pegasus-statistics [-h|--help]
                   [-o|--output dir]
                   [-c|--conf propfile]
                   [-p|--statistics-level level]
                   [-t|--time-filter filter]
                   [-i|--ignore-db-inconsistency]
                   [-v|--verbose]
                   [-q|--quiet]
                   [-m|--multiple-wf]
                   [-p|--ispmc]
                   [-u|--isuuid]
                   [[submitdir ..] | [workflow_uuid ..]]

Description

pegasus-statistics generates statistics about the workflow run like total jobs/tasks/sub workflows ran, how many succeeded/failed etc. It generates job instance statistics like run time, condor queue delay etc. It generates invocation statistics information grouped by transformation name. It also generates job instance and invocation statistics information grouped by time and host.

Options

-h , --help
Prints a usage summary with all the available command-line options.
-o dir , --output dir
Writes the output to the given directory.
-c propfile , --conf propfile
The properties file to use. This option overrides all other property files.
-s level , --statistics-level level

Specifies the statistics information to generate. Valid levels are: all, summary, wf_stats, jb_stats, tf_stats, ti_stats, int_stats. Default is summary. The output generated by pegasus-statistics is based on the the level set:

  • all: generates all the statistics information.
  • summary: generates the workflow statistics summary. In the case of a hierarchical workflow the summary is across all sub workflows.
  • wf_stats: generates the workflow statistics information of each individual workflow. In case of a hierarchical workflow the workflow statistics are created for each sub workflow.
  • jb_stats: generates the job statistics information of each individual workflow. In case of hierarchical workflow the job statistics is created for each sub workflows. Note: Not supported when generating statistics over multiple workflows.
  • tf_stats: generates the invocation statistics information of each individual workflow grouped by transformation name .In case of hierarchical workflow the transformation statistics is created for each sub workflows.
  • ti_stats: generates the job instance and invocation statistics like total count and runtime grouped by time and host.
  • int_stats: generates integrity metrcis like count, duration of integrity checks.
-t filter , --time-filter filter
Specifies the time filter to group the time statistics. Valid filter values are: month, week, day, hour. Default is day.
-i , --ignore-db-inconsistency
Turn off the the check for database consistency.
-v , --verbose
Increases the log level. If omitted, the default level will be set to WARNING. When this option is given, the log level is changed to INFO. If this option is repeated, the log level will be changed to DEBUG.
-q , --quiet
Decreases the log level. If omitted, the default level will be set to WARNING. When this option is given, the log level is changed to ERROR.
-m , --multiple-wf
Set this option when generating statistics over more than one workflow. The tool automatically sets this flag if multiple submit directories or multiple workflow UUIDs are provided. This option would need to be set explicitly only to generate statistics over all workflows in a single STAMPEDE database. NOTE: When workflows are specified as UUIDs the --conf options needs to be set for the tool to determine the STAMPEDE database URL.
-p , --ispmc
Set this flag to generate statistics for workflows which are run with PMC clustering enabled. It is recommended that this option be used when calculating statistics over multiple workflow runs.
-u , --isuuid
Set this option if the positional argument are workflow UUIDs. NOTE: When workflows are specified as UUIDs the --conf options needs to be set for the tool to determine the STAMPEDE database URL.

Example

Runs pegasus-statistics and writes the output to the given directory:

$ pegasus-statistics  -o /scratch/statistics /scratch/grid-setup/run0001

Runs pegasus-statistics over a workflow run identified by a single workflow UUID:

$ pegasus-statistics  --conf pegasusrc --isuuid 316f2986-7754-44ec-8b38-fcd0cb602ce0

Runs pegasus-statistics over a workflow run identified by a multiple workflow UUID:

$ pegasus-statistics  --conf pegasusrc --isuuid 316f2986-7754-44ec-8b38-fcd0cb602ce0 \
7ef77af8-4eb2-45ca-b37d-c5a02186133a

Runs pegasus-statistics over all workflows in the STAMPEDE database:

$ pegasus-statistics  --conf pegasusrc --multiple-wf

Authors

Prasanth Thomas Rajiv Mayani

Pegasus Team http://pegasus.isi.edu