Implementation:Apache Airflow Helm Template Deployment

Knowledge Sources	Apache Airflow
Domains	Kubernetes, Helm
Last Updated	2026-02-08 00:00 GMT

Overview

Concrete tool for deploying and scaling Airflow components on Kubernetes provided by the Helm chart templates.

Description

The Helm chart contains 103 template files that generate Kubernetes resources for all Airflow components. Key templates include scheduler-deployment.yaml, worker-deployment.yaml, and associated Service, ConfigMap, and HPA templates.

Usage

Deploy via helm install/upgrade. Customize replicas, resources, and scaling via values.yaml.

Code Reference

Source Location

• Repository: Apache Airflow
• File: chart/templates/ (103 template files)
• File: chart/Chart.yaml (chart metadata, apiVersion v2, appVersion 3.1.7)

Signature

# Key component configuration
scheduler:
  replicas: 1
  resources: {}
  # livenessProbe, readinessProbe, nodeSelector, tolerations, affinity

webserver:
  replicas: 1
  resources: {}

workers:
  replicas: 1
  persistence:
    enabled: true
  hpa:
    enabled: false
    minReplicaCount: 1
    maxReplicaCount: 5

triggerer:
  replicas: 1

dagProcessor:
  replicas: 1

Import

# Deploy
helm install airflow apache-airflow/airflow -f my-values.yaml

# Scale workers
kubectl scale deployment airflow-worker --replicas=5

# Or via HPA
helm upgrade airflow apache-airflow/airflow \
  --set workers.hpa.enabled=true \
  --set workers.hpa.maxReplicaCount=10

I/O Contract

Inputs

Name	Type	Required	Description
Component replicas	int	No	Number of replicas per component
Component resources	dict	No	CPU/memory requests and limits
HPA settings	dict	No	Autoscaling configuration
Kubernetes cluster	Cluster	Yes	Target cluster

Outputs

Name	Type	Description
Deployments	K8s Deployments	Running Airflow component pods
Services	K8s Services	Internal service endpoints
Ingress	K8s Ingress	External access to webserver/API
HPA	K8s HPA	Auto-scaling policies

Usage Examples

Production Scaling

# my-values.yaml
scheduler:
  replicas: 2
  resources:
    requests:
      cpu: "1"
      memory: "2Gi"

workers:
  replicas: 3
  hpa:
    enabled: true
    minReplicaCount: 3
    maxReplicaCount: 20
    metrics:
      - type: Resource
        resource:
          name: cpu
          target:
            type: Utilization
            averageUtilization: 80

Related Pages

Implements Principle

• Principle:Apache_Airflow_Component_Deployment_Scaling

Requires Environment

• Environment:Apache_Airflow_Kubernetes_Helm_Environment