Laravel queue jobs not processing: a checklist that finds the cause
Jobs queue but never run. Work through worker, connection, failed-job and deployment checks in this order to find out why.
If a job is queued and never executed, either no worker is consuming the queue, the worker is consuming a different queue, or the job fails instantly and lands in failed_jobs. Each has a distinct symptom.
Confirm a worker is running and in the right environment
In a terminal, php artisan queue:work should print processing lines as jobs arrive. On a server, use Supervisor or Horizon. A worker started before a deploy is still running the old code, so restart workers as part of every release: php artisan queue:restart.
- Check the process list for a live queue:work or horizon process
- Make sure the worker is on the same QUEUE_CONNECTION as the app
- Restart workers after deploys with php artisan queue:restart
Check which queue the job was pushed to
A job dispatched with onQueue("emails") sits in the emails queue. A worker started without --queue=emails only consumes default, so the job waits forever. Either pass the queue name to the worker or drop the custom queue name.
Look in failed_jobs before anything else
php artisan queue:failed lists attempts, exceptions and payloads. A job that fails immediately pushes itself to failed_jobs after its retry limit, so "nothing happens" often means "it failed three times in two seconds". Read the exception — usually a missing class, a serialisation problem or a database lock.
Serialisation and the model it references
Jobs receive serialised payloads. If a job takes a model ID and that row is deleted before the worker runs, the job throws ModelNotFoundException. Use $job->delete() semantics deliberately, or catch the exception and treat it as done. Class properties that cannot be serialised (closures, PDO handles) fail on push.
Timeouts, memory and the retry window
A job that exceeds the worker timeout is killed and retried, which can look like it never runs on a long task. Match the job timeout to the worker timeout, raise memory limits for large payloads, and check that retry_after is longer than your longest job or the same job will be processed twice.
Frequently asked questions
Why do jobs work locally but not on the server?
The server usually has no worker process, or Supervisor is running with an old release path. Confirm the process exists and that its working directory points at the current release.
Should I use sync in production?
No. The sync driver runs jobs inline, which hides queue problems and makes slow requests. It is fine in tests; in production use database or Redis with a supervised worker.
How many workers do I need?
Start with one and watch queue wait time. Add workers when latency grows, but remember each worker is a PHP process with its own memory; a database connection pool of 10 will not support 20 workers.
More on this topic: Laravel articles.
Admin
Related articles
Upgrading Laravel 10 to 12: a realistic one-week checklist
Two major versions at once is mostly mechanical: dependency updates, the new bootstrap file, and a test suite that tells you what...
Livewire or Inertia with Vue/React: which is cheaper to build and keep?
Livewire keeps one language and one deploy; Inertia buys a real component ecosystem at the cost of a second build pipeline. The ch...