Your own command-line tools.
WHAT THEY ARE FOR
Maintenance tasks Data imports and exports One-off operations Scheduled work
WHY A COMMAND RATHER THAN A SCRIPT
It runs within the application, with access to models, configuration and services.
WHAT TO DEFINE
A name, a description, and any arguments and options.
WHAT TO PROVIDE
Clear output describing what is happening A confirmation prompt for anything destructive A mode that reports without acting
WHY THAT LAST ONE
It allows verification before committing.
WHAT TO RETURN
A meaningful exit status, so failures are detectable.
WHAT TO LOG
That it ran, what it did, and any failure.
WHAT TO BUILD IN
Safety for repeated runs Bounds on how much a single run may affect
WHAT TO BE CAREFUL WITH
Memory, when processing many records.
WHAT TO USE
Chunked processing, rather than loading everything.
WHAT TO TEST
The command, including its failure paths.
WHAT TO SCHEDULE
Anything recurring, through the scheduler rather than separate cron entries.
