BullMQ Trigger
AI-generatedSummary
Starts a BullMQ Worker on a named Redis queue and fires the workflow once for each job it picks up, emitting the job's details as output.
Inputs
- Queue Name (required) — Name of the BullMQ queue the worker listens to.
- Options > Concurrency — Number of jobs the worker may process concurrently. Passed to the worker only if set; the field default is 1.
- Options > Lock Duration — Time in milliseconds the job lock is held. Passed to the worker only if set; the field default is 30000.
- Redis credential (required) — The 'redis' credential. Its host, port, password and db are used for the connection.
Output shape
One workflow item per job, emitted on the main output. Each item has id, name, data (the job payload), opts, progress, attemptsMade and timestamp.
The worker handler only emits the item and does not return a result, so the job is marked completed once the emit call returns. It does not wait for the workflow to finish, and workflow failures do not fail the job. Worker errors are only logged to the console. The worker is closed when the workflow is deactivated. The node consumes jobs rather than passively observing them, so other workers on the same queue will compete for them.
Examples
Example 1: Process email-sending jobs pushed to a queue by a backend service
Queue Name = 'emails', Redis credential configured, options left empty; use job data fields downstream.
Example 2: Handle several jobs in parallel with longer-running steps
Queue Name = 'reports', Concurrency = 5, Lock Duration = 60000.