Introduction

If you’re using n8n in production and need horizontal scaling for better performance and reliability, queue mode is essential. In this guide, we’ll walk through deploying n8n on Kubernetes using the official Helm chart, configuring Redis as the central queue, and separating main and worker pods for distributed execution.

This setup:

Uses Helm to deploy n8n in queue mode
Runs main and worker pods separately
Integrates with PostgreSQL and Redis
Supports horizontal scalability with multiple workers

Architecture Overview

Browser UI --> Main Pod (Web UI + Scheduler)
                        |
                        v
                Redis Queue (Bull)
                        |
                        v
                Worker Pods (process jobs)

Prerequisites

Kubernetes cluster (tested with Minikube)
Helm v3+
Redis and PostgreSQL deployed (e.g., Bitnami Helm charts, AWS RDS, Elasticache)

Helm Values Breakdown

Here is the values.yaml you can use to deploy a queue-enabled n8n cluster:

Global Image Config

image:
  repository: n8nio/n8n
  pullPolicy: IfNotPresent
  tag: "1.102.4"

Scaling Config

scaling:
  enabled: true
  redis:
    host: redis-master.default.svc.cluster.local
    port: 6379

This enables queue mode and points n8n to the Redis instance handling the Bull queue.

Main Pod

The main pod handles the Web UI and workflow triggering (e.g., cron/schedule triggers):

main:
  config:
    n8n:
      basicAuth:
        active: true
        user: admin
        password: admin
      encryption_key: "your-random-secret-key"
      port: 5678
      protocol: http
    db:
      type: postgresdb
      postgresdb:
        host: postgres-postgresql.default.svc.cluster.local
        port: 5432
        database: n8n
        user: n8n
    queue:
      bull:
        redis:
          host: redis-master.default.svc.cluster.local
          port: 6379

  secret:
    db:
      postgresdb:
        password: secret

  extraEnv:
    EXECUTIONS_MODE:
      value: "queue"                      # Ensures execution uses Redis queue
    N8N_LOG_LEVEL:
      value: "debug"                     # Enables debug logs
    N8N_RUNNERS_ENABLED:
      value: "true"                      # Enables worker communication
    N8N_REDIS_TIMEOUT_THRESHOLD:
      value: "30000"
    N8N_ENFORCE_SETTINGS_FILE_PERMISSIONS:
      value: "false"
    QUEUE_HEALTH_CHECK_ACTIVE:
      value: "true"
    N8N_METRICS:
      value: "true"
    N8N_DIAGNOSTICS_ENABLED:
      value: "false"

Worker Pods

Workers are responsible for consuming and processing jobs from Redis:

worker:
  enabled: true
  replicaCount: 4

  config:
    db:
      type: postgresdb
      postgresdb:
        host: postgres-postgresql.default.svc.cluster.local
        port: 5432
        database: n8n
        user: n8n
    queue:
      bull:
        redis:
          host: redis-master.default.svc.cluster.local
          port: 6379

  secret:
    db:
      postgresdb:
        password: secret

  extraEnv:
    EXECUTIONS_MODE:
      value: "queue"                      # Must match main pod
    N8N_LOG_LEVEL:
      value: "debug"
    N8N_RUNNERS_ENABLED:
      value: "true"
    N8N_REDIS_TIMEOUT_THRESHOLD:
      value: "30000"
    N8N_ENFORCE_SETTINGS_FILE_PERMISSIONS:
      value: "false"
    QUEUE_HEALTH_CHECK_ACTIVE:
      value: "true"
    N8N_METRICS:
      value: "true"
    N8N_DIAGNOSTICS_ENABLED:
      value: "false"

  concurrency: 5

concurrency defines how many jobs each worker can process simultaneously.

Webhook Pod (Optional)

If you want to scale webhook handling independently:

webhook:
  enabled: false  # Enable if needed

Deployment

Save the above config to values.yaml
Deploy Redis & Postgres (optional):

helm install redis bitnami/redis
helm install postgres bitnami/postgresql

Deploy n8n:

helm upgrade --install n8n oci://8gears.container-registry.com/library/n8n --namespace n8n --create-namespace -f values.yaml

Verifying Worker Pods Are Running

After deployment, ensure that your worker pods are up and running. You can verify this by running:

kubectl get pods -n n8n

You should see multiple n8n-worker pods in the Running state, like this:

Each of these workers is connected to the Redis queue and ready to pick up jobs. The number of workers is controlled via the worker.replicaCount value in your values.yaml. In our case, we set it to 4, and Kubernetes has spawned four workers accordingly.

You’ll be greeted with an owner account setup screen like the one below:

Here, you need to provide:

A valid email
First and last name
A strong password (at least 8 characters, 1 number, and 1 capital letter)

This setup screen only appears on first run. The credentials you set here will be stored in the database (users table in Postgres).

This step finalizes your n8n instance and gives you admin access to the system.

Closing Thoughts

Deploying n8n on Kubernetes using queue mode unlocks true scalability for production-grade workflows. By decoupling execution from the UI and enabling horizontal scaling with worker pods, you gain better control over performance, reliability, and system health.

Whether you’re automating internal ops or building external automation services, this Helm-based setup gives you a powerful, cloud-native foundation.

If you’re using Minikube for local testing or managing large workloads in production — this architecture holds up with minimal configuration drift between environments.

Now that your n8n setup is:

✅ Running in queue mode
✅ Using Redis-backed job queues
✅ Horizontally scalable via worker replicas
✅ Secure with basic auth
✅ Observable via logs and metrics

You’re ready to build faster and scale smarter.

Full `values.yaml` Reference

Below is the complete values.yaml file used for this setup. You can copy-paste it and adjust based on your Redis/Postgres service names, resource requirements, or any other preferences.

image:
  repository: n8nio/n8n
  pullPolicy: IfNotPresent
  tag: "1.102.4"
imagePullSecrets: []

nameOverride:
fullnameOverride:

hostAliases: []

ingress:
  enabled: false
  annotations: {}
  className: ""
  hosts:
    - host: workflow.example.com
      paths: []
  tls: []

scaling:
  enabled: true
  redis:
    host: redis-master.default.svc.cluster.local
    port: 6379

main:
  config:
    n8n:
      basicAuth:
        active: true
        user: admin
        password: admin
      encryption_key: "random-generated-secret-key"
      port: 5678
      protocol: http
    db:
      type: postgresdb
      postgresdb:
        host: postgres-postgresql.default.svc.cluster.local
        port: 5432
        database: n8n
        user: n8n
    queue:
      bull:
        redis:
          host: redis-master.default.svc.cluster.local
          port: 6379
  secret:
    db:
      postgresdb:
        password: secret

  extraEnv:
    EXECUTIONS_MODE: 
      value: "queue"
    N8N_LOG_LEVEL:
      value: "debug"
    N8N_RUNNERS_ENABLED:
      value: "true"
    N8N_REDIS_TIMEOUT_THRESHOLD:
      value: "30000"
    N8N_ENFORCE_SETTINGS_FILE_PERMISSIONS:
      value: "false"
    QUEUE_HEALTH_CHECK_ACTIVE:
      value: "true"
    N8N_METRICS:
      value: "true"
    N8N_DIAGNOSTICS_ENABLED:
      value: "false"

  persistence:
    enabled: false
    type: emptyDir
    accessModes:
      - ReadWriteOnce
    size: 1Gi

  extraVolumes: []

  extraVolumeMounts: []

  replicaCount: 1

  deploymentStrategy:
    type: "Recreate"
    #  maxSurge: "50%"
    #  maxUnavailable: "50%"

  serviceAccount:
    create: true
    annotations: {}
    name: ""

  deploymentAnnotations: {}

  deploymentLabels: {}

  podAnnotations: {}

  podLabels: {}

  podSecurityContext:
    runAsNonRoot: true
    runAsUser: 1000
    runAsGroup: 1000
    fsGroup: 1000

  securityContext: {}

  lifecycle: {}

  command: []

  livenessProbe:
    httpGet:
      path: /healthz
      port: 5678
    initialDelaySeconds: 30
    periodSeconds: 10
    timeoutSeconds: 5
    failureThreshold: 6
    successThreshold: 1

  readinessProbe:
    httpGet:
      path: /healthz
      port: 5678
    initialDelaySeconds: 30
    periodSeconds: 10
    timeoutSeconds: 5
    failureThreshold: 6
    successThreshold: 1

  initContainers: []

  service:
    enabled: true
    annotations: {}
    type: ClusterIP
    port: 80

  resources:
    limits:
      cpu: 200m
      memory: 512Mi
    requests:
      cpu: 200m
      memory: 512Mi

  autoscaling:
    enabled: false
    minReplicas: 1
    maxReplicas: 100
    targetCPUUtilizationPercentage: 80
    # targetMemoryUtilizationPercentage: 80

  nodeSelector: {}
  tolerations: []
  affinity: {}

worker:
  enabled: true
        
  config:
    db:
      type: postgresdb
      postgresdb:
        host: postgres-postgresql.default.svc.cluster.local
        port: 5432
        database: n8n
        user: n8n
    queue:
      bull:
        redis:
          host: redis-master.default.svc.cluster.local
          port: 6379
  secret:
    db:
      postgresdb:
        password: secret
  extraEnv:
    EXECUTIONS_MODE: 
      value: "queue"
    N8N_LOG_LEVEL:
      value: "debug"
    N8N_RUNNERS_ENABLED:
      value: "true"
    N8N_REDIS_TIMEOUT_THRESHOLD:
      value: "30000"
    N8N_ENFORCE_SETTINGS_FILE_PERMISSIONS:
      value: "false"
    QUEUE_HEALTH_CHECK_ACTIVE:
      value: "true"
    N8N_METRICS:
      value: "true"
    N8N_DIAGNOSTICS_ENABLED:
      value: "false"
  concurrency: 5

  persistence:
    enabled: false
    type: emptyDir
    accessModes:
      - ReadWriteOnce
    size: 1Gi

  replicaCount: 4

  deploymentStrategy:
    type: "Recreate"
    # maxSurge: "50%"
    # maxUnavailable: "50%"

  serviceAccount:
    create: true
    annotations: {}
    name: ""

  deploymentAnnotations: {}

  deploymentLabels: {}

  podAnnotations: {}

  podLabels: {}

  podSecurityContext:
    runAsNonRoot: true
    runAsUser: 1000
    runAsGroup: 1000
    fsGroup: 1000

  securityContext: {}

  lifecycle: {}

  command: []

  commandArgs: []

  livenessProbe:
    httpGet:
      path: /healthz
      port: http
    initialDelaySeconds: 30
    periodSeconds: 10
    timeoutSeconds: 5
    failureThreshold: 6
    successThreshold: 1

  readinessProbe:
    httpGet:
      path: /healthz
      port: http
    initialDelaySeconds: 30
    periodSeconds: 10
    timeoutSeconds: 5
    failureThreshold: 6
    successThreshold: 1

  initContainers: []

  service:
    annotations: {}
    type: ClusterIP
    port: 80

  resources:
    limits:
      cpu: 200m
      memory: 512Mi
    requests:
      cpu: 200m
      memory: 512Mi

  autoscaling:
    enabled: false
    minReplicas: 1
    maxReplicas: 100
    targetCPUUtilizationPercentage: 80
    # targetMemoryUtilizationPercentage: 80

  nodeSelector: {}
  tolerations: []
  affinity: {}

webhook:
  enabled: false

  config: {}

  secret: {}

  extraEnv:
    N8N_RUNNERS_ENABLED:
      value: "true"
    N8N_REDIS_TIMEOUT_THRESHOLD:
      value: "30000"
    N8N_ENFORCE_SETTINGS_FILE_PERMISSIONS:
      value: "false"
  #   WEBHOOK_URL:
  #   value: "http://webhook.domain.tld"


  persistence:
    enabled: false
    type: emptyDir
    accessModes:
      - ReadWriteOnce
    size: 1Gi

  replicaCount: 1

  deploymentStrategy:
    type: "Recreate"

  nameOverride: ""
  fullnameOverride: ""

  serviceAccount:
    create: true
    annotations: {}
    name: ""

  deploymentAnnotations: {}
  deploymentLabels: {}
  podAnnotations: {}
  podLabels: {}

  podSecurityContext:
    runAsNonRoot: true
    runAsUser: 1000
    runAsGroup: 1000
    fsGroup: 1000

  securityContext: {}

  lifecycle: {}

  command: []

  commandArgs: []

  livenessProbe:
    httpGet:
      path: /healthz
      port: http
    initialDelaySeconds: 30
    periodSeconds: 10
    timeoutSeconds: 5
    failureThreshold: 6
    successThreshold: 1

  readinessProbe:
    httpGet:
      path: /healthz
      port: http
    initialDelaySeconds: 30
    periodSeconds: 10
    timeoutSeconds: 5
    failureThreshold: 6
    successThreshold: 1

  initContainers: []

  service:
    annotations: {}
    type: ClusterIP
    port: 80

  resources: {}
  autoscaling:
    enabled: false
    minReplicas: 1
    maxReplicas: 100
    targetCPUUtilizationPercentage: 80
    # targetMemoryUtilizationPercentage: 80
  nodeSelector: {}
  tolerations: []
  affinity: {}

extraManifests: []

extraTemplateManifests: []

valkey:
  enabled: false

n8n on K8s: Scalable Worker Setup with Redis Queue

Introduction

Architecture Overview

Prerequisites

Helm Values Breakdown

Global Image Config

Scaling Config

Main Pod

Worker Pods

Webhook Pod (Optional)

Deployment

Verifying Worker Pods Are Running

Closing Thoughts

Full `values.yaml` Reference

Resources

Company

Contact

London

Istanbul

n8n on K8s: Scalable Worker Setup with Redis Queue

Introduction

Architecture Overview

Prerequisites

Helm Values Breakdown

Global Image Config

Scaling Config

Main Pod

Worker Pods

Webhook Pod (Optional)

Deployment

Verifying Worker Pods Are Running

Closing Thoughts

Full values.yaml Reference

Resources

Company

Contact

London

Istanbul

Full `values.yaml` Reference