AutoSys: List Only Failed Jobs with autorep

Thousands of jobs ran overnight and you need the failures. autorep lists everything; a grep on the status column gives you just the failures. Remember TE too.

Filter on Status

autorep -J ALL prints one line per job with its status. grep for the status with spaces around it.

autorep -J ALL | grep ' FA '

Include Terminated

TE means the job was killed: term_run_time, KILLJOB or a box terminator. It's a failure for operations.

autorep -J ALL | grep -E ' (FA|TE) '

Narrow, Count, Detail

Use a wildcard for one application, -c for a count, and -d for one job's run history.

autorep -J FIN% | grep -E ' (FA|TE) '
autorep -J ALL | grep -cE ' (FA|TE) '
autorep -J GL_POST -d

A Morning Check

Run it from cron or as an AutoSys job and alert when the count isn't zero. For important jobs, alarm_if_fail: 1 raises an alarm the moment they fail.

#!/bin/sh
n=$(autorep -J ALL | grep -cE ' (FA|TE) ')
if [ "$n" -gt 0 ]; then
  autorep -J ALL | grep -E ' (FA|TE) ' | mail -s "AutoSys: $n failed" oncall@example.com
fi