CloudDeploy流水线触发报错：Python入口点未识别（已配置Procfile）

用Cloud Build + Cloud Deploy自动化GKE部署时，触发流水线后报以下错误：

Failed to run /bin/build: for Python, an entrypoint must be manually set, either with "GOOGLE_ENTRYPOINT" env var or by creating a "Procfile" file"

已在代码根目录（与cloudbuild.yaml、clouddeploy.yaml同目录）创建Procfile，内容为：

web: gunicorn --bind :8080 main:app

相关配置文件如下：

cloudbuild.yaml #

steps:
- name: 'gcr.io/k8s-skaffold/pack'
  entrypoint: 'pack'
  args: ['build', '--builder=gcr.io/buildpacks/builder', '--publish', 'us-east1-docker.pkg.dev/fakeproject/my-docker-repo/app3:$SHORT_SHA']
- name: 'gcr.io/google.com/cloudsdktool/cloud-sdk'
  entrypoint: 'bash'
  args:
  - '-c'
  - > 
    gcloud deploy releases create release-$BUILD_ID
    --delivery-pipeline=app3
    --region=us-east1
    --source=./
    --images=app3=us-east1-docker.pkg.dev/fakeproject/my-docker-repo/app3:$SHORT_SHA
options:
  logging: CLOUD_LOGGING_ONLY

clouddeploy.yaml #

---
apiVersion: deploy.cloud.google.com/v1beta1
kind: DeliveryPipeline
metadata:
  name: app3
description: App3 Deployment Pipeline 
serialPipeline:
  stages:
  - targetId: demo-gke
---
apiVersion: deploy.cloud.google.com/v1beta1
kind: Target
metadata:
  name: demo-gke
description: Demo Environment
gke:
  cluster: projects/fakeproject/locations/us-east1/clusters/demo-gke
executionConfigs:
- usages:
  - DEPLOY
  workerPool: "projects/fakeproject/locations/us-east1/workerPools/cloud-build-pool"
- usages:
  - RENDER
  - VERIFY

main.py #

from flask import Flask
app = Flask(__name__)
@app.route("/")
def hello():
    return "Hello from app3"
# run the app.
if __name__ == "__main__":
    app.run()

requirements.txt和skaffold.yaml与参考教程完全一致。

• 严格校验Procfile的命名与位置：必须用大写P开头的Procfile（不能是小写procfile或其他变体），且必须放在Cloud Build构建使用的源代码根目录下。Google Cloud Buildpacks对文件名大小写敏感，命名错误会直接忽略该文件。
• 直接在pack命令中强制指定入口点：如果Procfile仍未被识别，无需纠结，直接修改cloudbuild.yaml的第一个步骤，通过环境变量硬设入口点：

  steps:
  - name: 'gcr.io/k8s-skaffold/pack'
    entrypoint: 'pack'
    args: [
      'build', 
      '--builder=gcr.io/buildpacks/builder', 
      '--publish', 'us-east1-docker.pkg.dev/fakeproject/my-docker-repo/app3:$SHORT_SHA',
      '--env', 'GOOGLE_ENTRYPOINT="gunicorn --bind :8080 main:app"'
    ]
  

• 检查Cloud Build的源代码上下文：确认触发器配置的代码分支中确实包含Procfile，且未被.gitignore或构建排除规则过滤，避免拉取错误分支或遗漏文件。
• 排查skaffold.yaml的覆盖配置：如果skaffold.yaml中定义了command或args字段，会直接覆盖Buildpacks设置的入口点。要么删除这些配置，要么修改为与Procfile一致的指令。
• 确认requirements.txt包含gunicorn依赖：确保requirements.txt中添加了gunicorn，避免后续容器启动时因依赖缺失报错。

内容的提问来源于stack exchange，提问作者Gary Turner