What is n8n queue mode?

It is a mode where one main process queues executions in Redis and separate worker processes run them, so many workflows can run in parallel.

Does queue mode need Postgres?

Yes. All processes must share one database, so Postgres is used instead of SQLite.

Why do my workers fail to run workflows?

A common cause is a different encryption key on the workers, or workers that cannot reach Redis or the database.

n8n Queue Mode Explained: Redis, Workers & Webhooks

n8n Queue Mode Explained: Redis, Workers and Webhook Processors

26 October 2026
by Admin

Copied successfully! The link has been copied to clipboard

n8n Queue Mode Explained: Redis, Workers and Webhook Processors

By default n8n runs in one process. It serves the editor, receives webhooks and runs every workflow itself. That is fine for most people. When you run many workflows at once, or a few heavy ones block the rest, queue mode spreads the work across several processes.

How queue mode works

Main instance. Serves the editor, handles schedules and puts executions into a queue.
Redis. Holds the queue of waiting executions.
Workers. Separate n8n processes that take jobs from the queue and run them. You can start more workers as load grows.
Webhook processors (optional). Separate processes that receive incoming webhook requests, so a busy main instance does not slow them down.
Postgres. Every process shares the same database.

Do you need it?

Signs that queue mode may help: executions wait a long time to start, the editor becomes slow while workflows run, or a heavy workflow stops others. If you only have a few light workflows, do not add this complexity. First try a bigger server and pruning, as described in our performance tuning guide.

Requirements

Postgres instead of SQLite.
A Redis server that all processes can reach.
The same encryption key set on every process, otherwise workers cannot decrypt credentials.
Enough RAM for the main instance and each worker.

The key settings

# on main, workers and webhook processors
EXECUTIONS_MODE=queue
QUEUE_BULL_REDIS_HOST=redis
QUEUE_BULL_REDIS_PORT=6379
DB_TYPE=postgresdb
N8N_ENCRYPTION_KEY=the-same-key-everywhere

Start a worker with the worker command and a webhook processor with the webhook command, using the same image and environment. Variable names can vary by version, so check the official documentation for yours.

Common mistakes

Different encryption keys on different processes.
Workers cannot reach Redis or Postgres because of Docker networking or firewall rules.
Webhooks sent to the main instance URL while a webhook processor is expected, or the reverse. Make sure your proxy routes webhook paths correctly.
Adding workers on a server that has no spare RAM or CPU. More processes need more resources.

Sizing

Each worker is a full n8n process, so plan RAM for each. A VPS with 4 vCPU and 8 to 16 GB RAM is a reasonable start for a main instance and two workers. See which VPS plan fits which n8n workload.

If you want queue-style scaling without running Redis and workers yourself, talk to us about managed n8n hosting and your expected workload.

Not sure if you need queue mode? Tell our team your workload and we will suggest a setup.

Contact PCRDP

Frequently asked questions

What is n8n queue mode?
It is a mode where one main process queues executions in Redis and separate worker processes run them, so many workflows can run in parallel.
Does queue mode need Postgres?
Yes. All processes must share one database, so Postgres is used instead of SQLite.
Why do my workers fail to run workflows?
A common cause is a different encryption key on the workers, or workers that cannot reach Redis or the database.

n8n Queue Mode Explained: Redis, Workers and Webhook Processors