AKS部署Airflow查看DAG任务日志报错：缺失http/https协议

问题现象 #

在Azure Kubernetes Service（AKS）中通过Helm Chart（版本1.11.0）部署Airflow，使用Azure Postgres作为元数据库，运行DAG后打开任务日志时出现错误：

*** Could not read served logs: Request URL is missing an 'http://' or 'https://' protocol

部署相关信息 #

部署命令 #

# Add helm repo for airflow
helm repo add apache-airflow https://airflow.apache.org
# Install airflow helm chart version 1.11.0 with custom configurations 
helm upgrade airflow apache-airflow/airflow --version 1.11.0 --install --namespace <namespace> --create-namespace -f override-values.yaml

自定义配置文件（override-values.yaml） #

webserverSecretKey: "<webserver secret key created using official helm production guide>"
scheduler:
  waitForMigrations:
    enabled: false
initDb:
  enabled: true
# Disable postgre instance to be created
postgresql:
  enabled: false
# Update storage class name for logs pvc to be managed-premium
logs:
  persistence:
    storageClassName: managed-premium
# Update storage class name for dags pvc to be managed-premium
dags:
  persistence:
    storageClassName: managed-premium
# Update storage class name for triggerer pvc to be managed-premium
triggerer:
  persistence:
    storageClassName: managed-premium
# Update storage class name for workers pvc to be managed-premium
workers:
  persistence:
    storageClassName: managed-premium
# For Postgres azure connection
pgbouncer:
 enabled: true
pgbouncer:
  # The maximum number of connections to PgBouncer
  maxClientConn: 100
  # The maximum number of server connections to the metadata database from PgBouncer
  metadataPoolSize: 10
  # The maximum number of server connections to the result backend database from PgBouncer
  resultBackendPoolSize: 5
env:
  - name: AIRFLOW__CORE__TEST_CONNECTION
    value: Enabled
  - name: AIRFLOW__WEBSERVER__EXPOSE_CONFIG
    value: 'true'
  - name: AIRFLOW__API__AUTH_BACKENDS
    value: "airflow.api.auth.backend.basic_auth"
# For Postgres azure connection
data:
  metadataSecretName: <secret having azure postgres database connection string>
# For DAGs git sync
dags:
  gitSync:
    enabled: true
    repo: <repo url>
    branch: "<branch name in repo>"
    subPath: ""
    sshKeySecret: <secret holding ssh key for repo>

解决方案 #

该错误源于Airflow WebServer无法正确生成日志服务的请求URL，需通过以下步骤修复：

1. 添加WebServer基础URL配置
   在override-values.yaml的env节点中补充以下环境变量，指定WebServer的完整访问地址（带协议）：

   env:
     # 保留原有环境变量
     - name: AIRFLOW__WEBSERVER__BASE_URL
       value: "http://airflow-webserver.<namespace>.svc.cluster.local:8080"
     - name: AIRFLOW__LOGGING__REMOTE_LOGGING
       value: "False"
   

   说明：

   • airflow-webserver是Helm部署默认的WebServer服务名，若自定义过服务名需替换
   • <namespace>替换为实际部署Airflow的命名空间
   • 若通过Ingress对外暴露服务，可将BASE_URL设为外部访问域名（如https://airflow.yourdomain.com）

2. 验证日志存储PVC状态
   执行命令确认日志持久化存储PVC已正常绑定：

   kubectl get pvc -n <namespace>
   

   确保airflow-logs的PVC状态为Bound，若未绑定需检查AKS中managed-premium存储类是否正常可用。

3. 重新部署Airflow
   应用更新后的配置重新部署：

   helm upgrade airflow apache-airflow/airflow --version 1.11.0 --install --namespace <namespace> -f override-values.yaml
   

部署完成后重新运行DAG，即可正常查看任务日志。

内容的提问来源于stack exchange，提问作者daemon22