Run a queue worker
Configure a database-backed queue and process background jobs.
On this page
Prepare the queue
Loom supports database and sync queue drivers. The default is database. Configure the application database and apply migrations before starting a worker:
php loom migrate
The default table names are jobs, failed_jobs, and job_batches. A sync driver executes work in the requesting process; a database-backed queue needs a worker.
Start a worker
Run from the application root:
php loom queue:work --queue=default
The worker stays running and processes jobs from that queue. In production, run it under your service manager and arrange for restarts when you deploy application code.
Set resource limits
php loom queue:work --queue=default --sleep=3 --memory=128 --timeout=60 --retry-after=90
| Option | Meaning |
|---|---|
--sleep |
Seconds to sleep when no jobs are available |
--memory |
Memory limit in MiB |
--timeout |
Job timeout in seconds |
--retry-after |
Seconds before a reserved job can be reclaimed |
--max-jobs |
Exit after processing this many jobs; 0 is unlimited |
Keep the job timeout shorter than retry-after so a long-running reservation is not reclaimed while its worker is still processing it. Make jobs safe to retry: external effects can happen before a job fails.
Diagnose an idle worker
Check the active queue driver, database connection, and queue name. Confirm that the application dispatches jobs to the same queue the worker consumes. Read application logs when jobs fail.
Use CLI diagnostics to inspect the environment. Worker defaults are defined in queue configuration.