适用于 PowerShell 7.0 及以上版本，需要 Docker Desktop 或 Podman
在现代 DevOps 工作流中，容器已经成为应用部署的标准载体。无论是微服务架构、CI/CD 流水线，还是本地开发环境，容器的使用无处不在。然而，当容器数量从几个增长到几十个甚至上百个时，手动管理就变得既低效又容易出错。如何用脚本化的方式编排和管理这些容器，就成了每个运维工程师必须面对的课题。
PowerShell 凭借其强大的对象管道、丰富的模块生态以及与 .NET 的深度集成，为容器编排提供了一种独特的自动化思路。与 Bash 脚本相比，PowerShell 能够直接操作结构化数据（如 JSON、YAML），将 Docker CLI 的文本输出转化为可查询的对象，从而实现更精细、更可靠的容器生命周期管理。
本文将围绕三个典型场景展开：使用 PowerShell 动态生成 Docker Compose 配置并管理多服务生命周期；构建容器健康检查与自动恢复机制；以及实现多环境批量容器部署工具，支持蓝绿部署和快速回滚。
Docker Compose 管理 在日常开发中，我们经常需要根据不同的环境（开发、测试、生产）生成不同的 Compose 配置。手动维护多份 YAML 文件不仅繁琐，还容易导致配置漂移。下面的脚本展示了如何用 PowerShell 动态生成 Docker Compose 文件，并统一管理服务的启动、停止和状态查询。
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 function New-DockerComposeConfig param ( [Parameter (Mandatory )] [ValidateSet ('dev' , 'staging' , 'prod' )] [string ]$Environment , [int ]$ReplicaCount = 1 , [int ]$MemoryLimitMb = 512 ) $envConfig = @ { dev = @ { HostPort = 8080 ; Tag = 'latest' ; LogLevel = 'Debug' } staging = @ { HostPort = 8081 ; Tag = 'rc' ; LogLevel = 'Information' } prod = @ { HostPort = 80 ; Tag = 'stable' ; LogLevel = 'Warning' } } $cfg = $envConfig [$Environment ] $compose = @ { version = '3.8' services = @ { webapp = @ { image = "myapp/webapp:$ ($cfg .Tag)" ports = @ ("$ ($cfg .HostPort):80" ) environment = @ ( "ASPNETCORE_ENVIRONMENT=$Environment " "LOG_LEVEL=$ ($cfg .LogLevel)" ) deploy = @ { replicas = $ReplicaCount resources = @ { limits = @ { memory = "$ {MemoryLimitMb}M" } reservations = @ { memory = "$ ([Math]::Floor($MemoryLimitMb / 2))M" } } restart_policy = @ { condition = 'on-failure' max_attempts = 3 } } healthcheck = @ { test = @ ('CMD' , 'curl' , '-f' , 'http://localhost/health' ) interval = '30s' timeout = '10s' retries = 3 } volumes = @ ('./data:/app/data' ) networks = @ ('app-network' ) } redis = @ { image = 'redis:7-alpine' ports = @ ('6379:6379' ) volumes = @ ('redis-data:/data' ) networks = @ ('app-network' ) healthcheck = @ { test = @ ('CMD' , 'redis-cli' , 'ping' ) interval = '15s' timeout = '5s' retries = 3 } } } networks = @ { 'app-network' = @ { driver = 'bridge' } } volumes = @ { 'redis-data' = @ {} } } $compose | ConvertTo-Yaml | Set-Content "docker-compose.$Environment .yml" Write-Host "已生成 docker-compose.$Environment .yml" -ForegroundColor Green return "docker-compose.$Environment .yml" } function Invoke-DockerComposeLifecycle param ( [Parameter (Mandatory )] [string ]$ComposeFile , [ValidateSet ('up' , 'down' , 'status' )] [string ]$Action = 'up' ) switch ($Action ) { 'up' { Write-Host "启动服务: $ComposeFile " -ForegroundColor Cyan docker compose -f $ComposeFile up -d --remove-orphans Write-Host "服务已启动，等待健康检查..." -ForegroundColor Green Start-Sleep -Seconds 5 docker compose -f $ComposeFile ps } 'down' { Write-Host "停止服务: $ComposeFile " -ForegroundColor Yellow docker compose -f $ComposeFile down --volumes --remove-orphans Write-Host "服务已停止并清理" -ForegroundColor Green } 'status' { docker compose -f $ComposeFile ps Write-Host "`n--- 资源使用 ---" -ForegroundColor Cyan docker stats --no-stream --format 'table {{.Name}}\t{{.CPUPerc}}\t{{.MemUsage}}' } } } $composeFile = New-DockerComposeConfig -Environment dev -ReplicaCount 2 -MemoryLimitMb 256 Invoke-DockerComposeLifecycle -ComposeFile $composeFile -Action up
执行结果示例：
1 2 3 4 5 6 7 8 9 10 11 12 已生成 docker-compose.dev.yml 启动服务: docker-compose.dev.yml [+] Running 3/3 ✔ Network dev_app-network Created ✔ Container dev-redis-1 Started ✔ Container dev-webapp-1 Started ✔ Container dev-webapp-2 Started 服务已启动，等待健康检查... NAME IMAGE STATUS PORTS dev-webapp-1 myapp/webapp:latest Up 5 seconds 0.0.0.0:8080->80/tcp dev-webapp-2 myapp/webapp:latest Up 5 seconds 0.0.0.0:8080->80/tcp dev-redis-1 redis:7-alpine Up 5 seconds 0.0.0.0:6379->6379/tcp
容器健康检查与自动恢复 在生产环境中，容器可能会因为内存溢出、依赖服务不可用或网络抖动等原因意外退出。如果缺乏自动化的监控和恢复机制，服务中断往往会持续到人工介入才得以解决。下面的脚本实现了一套轻量级的容器健康巡检系统，能够自动检测异常容器、触发重启，并在资源使用接近阈值时发出告警。
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 function Get-ContainerHealthReport param ( [string ]$Filter = '' , [int ]$CpuThreshold = 80 , [int ]$MemoryThresholdPercent = 85 ) $containers = docker ps -a --format '{{.ID}}|{{.Names}}|{{.Status}}|{{.Image}}' | ForEach-Object { $parts = $_ -split '\|' [PSCustomObject ]@ { Id = $parts [0 ] Name = $parts [1 ] Status = $parts [2 ] Image = $parts [3 ] } } if ($Filter ) { $containers = $containers | Where-Object { $_ .Name -match $Filter } } foreach ($c in $containers ) { $inspect = docker inspect $c .Id | ConvertFrom-Json $state = $inspect .State $health = $state .Health $report = [PSCustomObject ]@ { Name = $c .Name Image = $c .Image Running = $state .Running Health = if ($health ) { $health .Status } else { 'N/A' } Restarting = $state .Restarting ExitCode = $state .ExitCode StartedAt = $state .StartedAt } $stats = docker stats --no-stream --format '{{.CPUPerc}}|{{.MemUsage}}|{{.MemPerc}}' $c .Id if ($stats ) { $statParts = $stats -split '\|' $report | Add-Member -NotePropertyName 'CpuPct' ` -NotePropertyValue ($statParts [0 ] -replace '%' , '' ).Trim() $report | Add-Member -NotePropertyName 'MemPct' ` -NotePropertyValue ($statParts [2 ] -replace '%' , '' ).Trim() } $report } } function Repair-UnhealthyContainers param ( [int ]$MaxRestartAttempts = 3 , [int ]$CooldownSeconds = 30 , [switch ]$DryRun ) $report = Get-ContainerHealthReport $unhealthy = $report | Where-Object { -not $_ .Running -or $_ .Health -eq 'unhealthy' } if (-not $unhealthy ) { Write-Host "所有容器状态正常" -ForegroundColor Green return } foreach ($c in $unhealthy ) { $action = if (-not $c .Running) { '启动' } else { '重启' } if ($DryRun ) { Write-Host "[模拟] 将$action 容器: $ ($c .Name) (状态: $ ($c .Status))" ` -ForegroundColor Yellow continue } Write-Host "$action 容器: $ ($c .Name)..." -ForegroundColor Cyan docker restart $c .Name Start-Sleep -Seconds $CooldownSeconds $newState = docker inspect $c .Name --format '{{.State.Running}}' if ($newState -eq 'true' ) { Write-Host " 容器 $ ($c .Name) 已恢复运行" -ForegroundColor Green } else { Write-Host " 容器 $ ($c .Name) 恢复失败，需要人工介入" -ForegroundColor Red } } } Get-ContainerHealthReport | Format-Table -AutoSize Repair-UnhealthyContainers -DryRun
执行结果示例：
1 2 3 4 5 6 7 8 Name Image Running Health Restarting ExitCode CpuPct MemPct ---- ----- ------- ------ ---------- -------- ------ ------ dev-webapp-1 myapp/webapp:latest True healthy False 0 2.3% 15.2% dev-webapp-2 myapp/webapp:latest True healthy False 0 1.8% 14.7% dev-redis-1 redis:7-alpine True healthy False 0 0.5% 8.1% test-api-1 myapp/api:rc False unhealthy False 137 N/A N/A [模拟] 将重启 容器: test-api-1 (状态: Exited (137) 2 minutes ago)
批量容器部署工具 蓝绿部署是一种经典的零停机发布策略，它通过维护两套完全相同的生产环境（蓝和绿），在发布新版本时将流量从旧环境切换到新环境，从而实现无缝升级。结合 PowerShell 的参数化能力，我们可以构建一个支持多环境配置、蓝绿部署和一键回滚的批量部署工具。
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 class ContainerDeployment [string ]$ProjectName [string ]$Environment [hashtable ]$Services = @ {} ContainerDeployment([string ]$ProjectName , [string ]$Environment ) { $this .ProjectName = $ProjectName $this .Environment = $Environment } [void ] AddService ([string ]$Name , [string ]$Image , [string []]$Ports ) { $this .Services[$Name ] = @ { Image = $Image Ports = $Ports CreatedAt = (Get-Date ).ToString('o' ) } } [string ] Deploy ([string ]$Slot ) { $slotPrefix = "$ ($this .ProjectName)-$Slot " Write-Host "部署到 $Slot 槽位: $slotPrefix " -ForegroundColor Cyan foreach ($svc in $this .Services.GetEnumerator()) { $containerName = "$slotPrefix -$ ($svc .Key)" $portMap = ($svc .Value.Ports | ForEach-Object { '-p' ; $_ }) -join ' ' Write-Host " 拉取镜像: $ ($svc .Value.Image)" -ForegroundColor Gray docker pull $svc .Value.Image | Out-Null $existing = docker ps -aq -f "name=$containerName " if ($existing ) { docker stop $existing | Out-Null docker rm $existing | Out-Null } $runCmd = "docker run -d --name $containerName $portMap $ ($svc .Value.Image)" $containerId = Invoke-Expression $runCmd Write-Host " 已启动: $containerName ($ ($containerId .Substring(0,12)))" ` -ForegroundColor Green } return $slotPrefix } [void ] SwitchTraffic ([string ]$NewSlot ) { $oldSlot = if ($NewSlot -eq 'blue' ) { 'green' } else { 'blue' } $oldPrefix = "$ ($this .ProjectName)-$oldSlot " $newPrefix = "$ ($this .ProjectName)-$NewSlot " Write-Host "`n切换流量: $oldSlot -> $NewSlot " -ForegroundColor Yellow $newContainers = docker ps -f "name=$newPrefix " --format '{{.Names}}' foreach ($name in $newContainers ) { $health = docker inspect $name --format '{{.State.Health.Status}}' if ($health -ne 'healthy' ) { Write-Host " 容器 $name 状态为 $health ，中止切换" -ForegroundColor Red return } } $oldContainers = docker ps -f "name=$oldPrefix " --format '{{.Names}}' foreach ($name in $oldContainers ) { Write-Host " 停止旧容器: $name " -ForegroundColor Gray docker stop $name | Out-Null } Write-Host "流量已切换到 $NewSlot 槽位" -ForegroundColor Green } [void ] Rollback () { $activeSlot = $this .GetActiveSlot() $targetSlot = if ($activeSlot -eq 'blue' ) { 'green' } else { 'blue' } Write-Host "回滚到 $targetSlot 槽位..." -ForegroundColor Yellow $stopped = docker ps -aq -f "status=exited" -f "name=$ ($this .ProjectName)-$targetSlot " foreach ($id in $stopped ) { docker start $id | Out-Null $name = docker inspect $id --format '{{.Name}}' | ForEach-Object { $_ -replace '^/' , '' } Write-Host " 已恢复: $name " -ForegroundColor Green } $active = docker ps -f "name=$ ($this .ProjectName)-$activeSlot " --format '{{.Names}}' foreach ($name in $active ) { docker stop $name | Out-Null } Write-Host "回滚完成，活跃槽位: $targetSlot " -ForegroundColor Green } hidden [string ] GetActiveSlot () { $blue = docker ps -q -f "name=$ ($this .ProjectName)-blue" 2 >$null $green = docker ps -q -f "name=$ ($this .ProjectName)-green" 2 >$null if ($blue ) { return 'blue' } if ($green ) { return 'green' } return 'none' } } $deploy = [ContainerDeployment ]::new('myapp' , 'prod' )$deploy .AddService('web' , 'myapp/web:stable' , @ ('80:80' ))$deploy .AddService('api' , 'myapp/api:stable' , @ ('8080:8080' ))$deploy .AddService('worker' , 'myapp/worker:stable' , @ ())$deploy .Deploy('blue' )$deploy .SwitchTraffic('blue' )
执行结果示例：
1 2 3 4 5 6 7 8 9 10 11 12 13 部署到 blue 槽位: myapp-blue 拉取镜像: myapp/web:stable 已启动: myapp-blue-web (a1b2c3d4e5f6) 拉取镜像: myapp/api:stable 已启动: myapp-blue-api (f6e5d4c3b2a1) 拉取镜像: myapp/worker:stable 已启动: myapp-blue-worker (1a2b3c4d5e6f) 切换流量: green -> blue 停止旧容器: myapp-green-web 停止旧容器: myapp-green-api 停止旧容器: myapp-green-worker 流量已切换到 blue 槽位
注意事项
Docker 权限 ：在 Linux 上运行时，确保当前用户已加入 docker 用户组，否则需要在命令前加 sudo。生产环境建议使用 Rootless Docker 以降低安全风险。
ConvertTo-Yaml 模块 ：生成 Compose 文件依赖 powershell-yaml 模块，可通过 Install-Module -Name powershell-yaml -Scope CurrentUser 安装。如果无法安装，也可以用 ConvertTo-Json 生成 JSON 格式的配置。
健康检查延迟 ：容器启动后需要一定时间才能通过健康检查，Start-Sleep 的等待时间应根据应用的启动速度调整，避免误判。建议在脚本中加入轮询逻辑，而不是简单的固定等待。
蓝绿部署数据一致性 ：蓝绿切换时数据库迁移是一个常见陷阱。如果新版本包含破坏性的数据库变更，需要确保迁移脚本兼容新旧版本，否则回滚将变得不可行。建议将数据库迁移与容器部署解耦。
资源限制 ：Docker Desktop 在 macOS 和 Windows 上的资源配额受虚拟机限制，脚本中设置的 memory 限制不能超过 Docker Desktop 分配的总内存。可通过 Docker Desktop 设置面板调整。
日志与监控 ：脚本中的 docker stats 只能获取实时快照数据。如需长期监控和历史数据查询，建议集成 Prometheus + Grafana 或 Docker 原生的 docker logs --since 进行日志聚合分析。