新手向：GCP Cloud Functions部署Go REST API及自动化指南

一、优化后的项目文件夹结构 #

按照Go项目规范+GCF部署要求调整结构，让每个函数拥有独立部署上下文，公共依赖统一管理：

Root Folder
│
├── cmd/
│   └── main.go          # 本地测试入口（保留原有逻辑）
├── internal/            # 私有公共依赖模块（仅本项目可导入）
│   ├── models/
│   │   └── models.go
│   ├── validations/
│   │   └── validations.go
│   └── db/
│       └── firestore.go
├── functions/           # 每个云函数对应独立子目录
│   ├── get-all-employees/
│   │   └── main.go      # 原getAllEmployees逻辑，导入internal模块
│   ├── create-employee/
│   │   └── main.go
│   ├── update-employee/
│   │   └── main.go
│   └── delete-employee/
│       └── main.go
├── terraform/           # Terraform自动化配置目录
│   ├── provider.tf
│   ├── variables.tf
│   ├── main.tf
│   └── outputs.tf
└── go.mod

核心调整：将每个云函数拆分为独立目录，公共代码放入internal（符合Go包管理规范，避免被外部意外导入），每个函数目录下的main.go作为该函数的入口文件。

二、使用gcloud命令部署函数 #

1. 单个函数部署命令 #

确保每个函数目录下的main.go正确导入公共模块（注意go.mod中的module路径要匹配），比如：

package p

import (
    "your-module-path/internal/models"
    "your-module-path/internal/validations"
    "your-module-path/internal/db"
    "net/http"
)

// GetAllEmployees 云函数HTTP触发器入口
func GetAllEmployees(w http.ResponseWriter, r *http.Request) {
    // 原有业务逻辑
}

执行部署命令（以gen2版本为例）：

gcloud functions deploy get-all-employees \
  --gen2 \
  --runtime go121 \
  --region us-central1 \
  --source ./functions/get-all-employees \
  --entry-point GetAllEmployees \
  --trigger-http \
  --allow-unauthenticated # 生产环境建议移除，改用IAM控制访问

2. 批量部署脚本 #

编写deploy-all.sh脚本批量部署所有函数：

#!/bin/bash

# 函数列表：格式为"函数名:入口函数名"
FUNCTIONS=(
  "get-all-employees:GetAllEmployees"
  "create-employee:CreateEmployee"
  "update-employee:UpdateEmployee"
  "delete-employee:DeleteEmployee"
)

REGION="us-central1"
RUNTIME="go121"

for func in "${FUNCTIONS[@]}"; do
  IFS=":" read -r NAME ENTRY <<< "$func"
  echo "开始部署函数：$NAME"
  gcloud functions deploy "$NAME" \
    --gen2 \
    --runtime "$RUNTIME" \
    --region "$REGION" \
    --source "./functions/$NAME" \
    --entry-point "$ENTRY" \
    --trigger-http
done

赋予执行权限后运行：chmod +x deploy-all.sh && ./deploy-all.sh

三、Terraform自动化部署方案 #

1. 基础配置文件 #

terraform/provider.tf（GCP提供者配置）

terraform {
  required_providers {
    google = {
      source  = "hashicorp/google"
      version = "~> 4.0"
    }
  }
}

provider "google" {
  project = var.gcp_project_id
  region  = var.gcp_region
}

terraform/variables.tf（定义可配置变量）

variable "gcp_project_id" {
  type        = string
  description = "你的GCP项目ID"
}

variable "gcp_region" {
  type        = string
  description = "云函数部署区域"
  default     = "us-central1"
}

variable "function_runtime" {
  type        = string
  description = "Go运行时版本"
  default     = "go121"
}

2. 函数资源定义 #

terraform/main.tf（核心部署逻辑）

# 创建存储桶用于存放函数源码包
resource "google_storage_bucket" "function_source_bucket" {
  name     = "${var.gcp_project_id}-function-sources"
  location = var.gcp_region
  uniform_bucket_level_access = true
}

# 打包并上传所有函数源码到存储桶
resource "null_resource" "package_and_upload_functions" {
  depends_on = [google_storage_bucket.function_source_bucket]

  provisioner "local-exec" {
    command = <<EOT
      # 打包每个函数目录
      zip -r functions/get-all-employees.zip ./functions/get-all-employees -x "*.git*"
      zip -r functions/create-employee.zip ./functions/create-employee -x "*.git*"
      zip -r functions/update-employee.zip ./functions/update-employee -x "*.git*"
      zip -r functions/delete-employee.zip ./functions/delete-employee -x "*.git*"

      # 上传到GCS存储桶
      gsutil cp functions/*.zip gs://${google_storage_bucket.function_source_bucket.name}/functions/
    EOT
  }
}

# 部署"获取所有员工"函数
resource "google_cloudfunctions2_function" "get_all_employees" {
  name        = "get-all-employees"
  location    = var.gcp_region
  description = "获取员工列表的云函数"

  build_config {
    runtime     = var.function_runtime
    entry_point = "GetAllEmployees"
    source {
      storage_source {
        bucket = google_storage_bucket.function_source_bucket.name
        object = "functions/get-all-employees.zip"
      }
    }
  }

  service_config {
    available_memory = "256MiB"
    timeout_seconds = 60
    ingress_settings = "ALLOW_ALL"
    all_traffic_on_latest_revision = true
  }

  event_trigger {
    trigger_region = var.gcp_region
    event_type     = "google.cloud.http"
    http_trigger {
      security_level = "SECURE_OPTIONAL"
    }
  }

  depends_on = [null_resource.package_and_upload_functions]
}

# 部署"创建员工"函数（其余update/delete函数逻辑类似，复制后修改名称和入口即可）
resource "google_cloudfunctions2_function" "create_employee" {
  name        = "create-employee"
  location    = var.gcp_region
  description = "创建新员工的云函数"

  build_config {
    runtime     = var.function_runtime
    entry_point = "CreateEmployee"
    source {
      storage_source {
        bucket = google_storage_bucket.function_source_bucket.name
        object = "functions/create-employee.zip"
      }
    }
  }

  service_config {
    available_memory = "256MiB"
    timeout_seconds = 60
    ingress_settings = "ALLOW_ALL"
    all_traffic_on_latest_revision = true
  }

  event_trigger {
    trigger_region = var.gcp_region
    event_type     = "google.cloud.http"
    http_trigger {
      security_level = "SECURE_OPTIONAL"
    }
  }

  depends_on = [null_resource.package_and_upload_functions]
}

3. 输出函数访问URL #

terraform/outputs.tf

output "get_all_employees_url" {
  value = google_cloudfunctions2_function.get_all_employees.service_config.uri
}

output "create_employee_url" {
  value = google_cloudfunctions2_function.create_employee.service_config.uri
}

# 其余函数URL输出同理添加

4. Terraform部署步骤 #

1. 初始化Terraform：cd terraform && terraform init
2. 预览部署计划：terraform plan -var="gcp_project_id=你的项目ID"
3. 执行部署：terraform apply -var="gcp_project_id=你的项目ID"

内容的提问来源于stack exchange，提问作者Shreyas Awankar