流水线基础

工作原理

bitbucket-pipelines.yml	仓库根目录的配置文件
Docker containers	每个步骤在独立容器中运行
Triggers	push、PR、tag、定时或手动触发
Build minutes	配额取决于套餐级别

启用流水线

# Repository Settings → Pipelines → Enable
# Add bitbucket-pipelines.yml to repo root
# First push triggers the pipeline

bitbucket-pipelines.yml

最简配置

image: node:20
pipelines:
  default:
    - step:
        script:
          - npm install
          - npm test

按分支定义流水线

pipelines:
  branches:
    main:
      - step:
          script:
            - npm run build
            - npm run deploy

Tag 与 Pull Request 流水线

pipelines:
  tags:
    'v*':
      - step:
          script:
            - npm run release
  pull-requests:
    '**':
      - step:
          script:
            - npm test

步骤

步骤选项

name	步骤显示名称
image	覆盖全局 Docker 镜像
script	要执行的 shell 命令列表
size	1x（4GB）或 2x（8GB）内存
max-time	超时分钟数（默认 120）
trigger	设为 manual 表示仅手动触发

并行步骤

- parallel:
    - step:
        name: "Lint"
        script:
          - npm run lint
    - step:
        name: "Test"
        script:
          - npm test

手动步骤

- step:
    name: "Deploy to Production"
    trigger: manual
    script:
      - ./deploy.sh prod

变量

变量类型

Repository variables	Settings → Pipelines → Variables
Deployment variables	作用域限于部署环境
Secured variables	加密，日志中脱敏显示
Pipeline variables	在 YAML 中内联定义

使用变量

pipelines:
  default:
    - step:
        script:
          - echo $MY_VAR
          - docker login -u $DOCKER_USER -p $DOCKER_PASS

内置变量

$BITBUCKET_COMMIT	完整 commit SHA
$BITBUCKET_BRANCH	分支名称
$BITBUCKET_TAG	Tag 名称（tag 流水线）
$BITBUCKET_BUILD_NUMBER	递增构建编号
$BITBUCKET_REPO_SLUG	仓库 slug

缓存

预定义缓存

- step:
    caches:
      - node        # ~/.npm
      - pip         # ~/.cache/pip
      - docker      # Docker layer cache
    script:
      - npm install
      - npm test

自定义缓存

definitions:
  caches:
    gradle: ~/.gradle/caches
    mylibs: vendor/libs
pipelines:
  default:
    - step:
        caches:
          - gradle
        script:
          - ./gradlew build

缓存行为

Duration	缓存 7 天后过期
Scope	同仓库所有流水线共享
Clear	Pipelines → Caches → Delete

产物

步骤间传递文件

- step:
    name: "Build"
    script:
      - npm run build
    artifacts:
      - dist/**
- step:
    name: "Deploy"
    script:
      - ls dist/    # artifacts available
      - ./deploy.sh

产物选项

artifacts	要传递的文件 glob 模式
Download	后续步骤自动可用
Max size	每步骤最大 1 GB
Retention	构建后保留 14 天

部署

部署环境

- step:
    name: "Deploy Staging"
    deployment: staging
    script:
      - ./deploy.sh staging
- step:
    name: "Deploy Production"
    deployment: production
    trigger: manual
    script:
      - ./deploy.sh prod

环境类型

test	测试环境
staging	预生产环境
production	线上环境，在仪表板中追踪

常用模式

Docker 构建与推送

- step:
    services:
      - docker
    script:
      - docker build -t myapp:$BITBUCKET_COMMIT .
      - docker login -u $DOCKER_USER -p $DOCKER_PASS
      - docker push myapp:$BITBUCKET_COMMIT

服务容器

definitions:
  services:
    postgres:
      image: postgres:16
      variables:
        POSTGRES_DB: testdb
        POSTGRES_PASSWORD: secret
pipelines:
  default:
    - step:
        services:
          - postgres
        script:
          - npm test

使用 Pipe 的条件步骤

- step:
    name: "Deploy to S3"
    script:
      - pipe: atlassian/aws-s3-deploy:1.1.0
        variables:
          AWS_ACCESS_KEY_ID: $AWS_KEY
          AWS_SECRET_ACCESS_KEY: $AWS_SECRET
          S3_BUCKET: my-bucket
          LOCAL_PATH: dist/

BITBUCKET PIPELINES 快速参考