Usage

Failed Job Email

Queue Monitor can send out an email as soon as a queue job has failed. This can be helpful to alert site admins of potential issues with queue jobs.

Nominate a User Group in Queue Monitor's settings, with each user in that group receiving an email notification when a queue job fails. You can also add individual email addresses for recipients who do not need to be Craft users. The content of this email is customisable via the Craft System Messages utility.

Auto-Restart Failed Jobs

You can also easily manage auto-restarting of failed queue jobs. By default, only one attempt is made for a queue job, and if it fails, no more attempts are made. Queue Monitor allows you to manage how many attempts a failed job should retry, along with setting the duration in seconds between retries.

Stalled Queue Monitoring

Queue Monitor can also alert you when jobs are waiting but nothing is processing them, such as when the queue worker has stopped. Turn on stalled queue monitoring in Queue Monitor's settings, choose how many minutes a job can wait before the queue counts as stalled, and nominate a User Group, email addresses or a webhook to notify. After an alert is sent, another isn't sent until the notification interval has passed.

The check runs from a console command rather than from the queue, so it still works when the queue worker itself has stopped. Run it from cron, for example every five minutes:

*/5 * * * * /path/to/project/craft queue-monitor/stalled/check

Choose a Retry Policy

A retry is useful when a job failed because a service was temporarily unavailable. It will not repair an invalid template or missing credentials. For example, a notification group can alert the site's maintainers while a short retry policy gives an interrupted connection time to recover.

Set the recipient group or email addresses and the retry settings in the control panel. When a job fails, compare its error and attempt count in Craft's queue with the received notification. Correct a persistent cause before retrying again, then confirm that the job completes. For jobs that send data to an external service, also check the receiving account: completion in Craft and the expected remote result are both part of the check.