OpenVMS System Management Guide, 2nd Edition

After you have identified performance objectives, you should regularly monitor whether application performance goals are being satisfied. On a weekly or monthly basis you should summarize response times and publish the results to users or management. Such a report also becomes a historical record of response times, which is immensely helpful in analyzing performance over a longer term.
You can monitor resource usage and response times for batch jobs simply by reviewing the resource-accounting information included at the end of the log file for each batch job. To ensure that this information is retained, include the /KEEP and /LOG qualifiers when jobs are submitted. By default, a log file is created in the job owner s default directory. The log file is given the same name as the command procedure that submitted it with a file extension of .LOG. For example:
$ <b class="bold">SUBMIT/KEEP/NOPRINT DB_UPD.COM</b> . . . $ <b class="bold">TYPE DB_UPD.LOG</b> . . . SYSTEM job terminated at 20-SEP-2003 10:28:10.17 Accounting information: Buffered I/O count: 1545 Peak Working set size: 912 Buffered I/O count: 1034 Peak page file size: 11134 Direct I/O count: 8512 Mounted Volume: ...