Schedule-it

Schedule-it keeps your site's scheduled jobs in the repository. The host runs one cron entry every minute, and that entry works out which jobs are due and runs them:

* * * * *  php craft schedule-it/run

The schedule is config/schedule-it.php, so it is reviewed like any other code, it is the same on every environment that deploys the commit, and a new environment gets every job the moment that one entry exists — instead of jobs living only in a hosting dashboard, invisible to the repository and missing wherever someone forgot to re-add them.

Write jobs in words or as cron expressions

Job::command('gc/run')->dailyAt('03:00') and ->weekdays()->at('08:30') sit beside plain cron expressions, with a description, a timezone or a per-environment switch where you need one. craft schedule-it/list shows every job, the expression it became and when it is next due.

Failures are reported, never swallowed

• Jobs run through craft, with an absolute path, because cron does not promise a working directory.
• Every job's stdout and stderr go to the log file, and are replayed to the runner's output for hosts whose task runner throws its filesystem away.
• schedule-it/run reports each job's exit code and exits non-zero if any failed.
• A misspelt key, a bad expression or an unknown timezone is an error, never a job that silently does not run.
• It refuses a lock that other servers cannot see, so a job never runs twice at once across a cluster.

Console-only: the web app has no route to the runner. Built on omnilight/yii2-scheduling, with the defaults that library needs under Craft already applied.