在AWS ECS Fargate上部署Apache DevLake的实践指南

Apache DevLake是一个开源的数据湖平台，用于收集、分析和可视化软件开发过程中的各种数据。本文将详细介绍如何将DevLake从Docker Compose部署迁移到AWS ECS Fargate环境。

容器与卷配置转换

在Docker Compose中，我们通常使用volumes指令来定义数据持久化。当迁移到ECS Fargate时，需要将这些配置转换为ECS任务定义中的相应部分。

MySQL存储卷配置

MySQL容器需要持久化存储数据库文件。在ECS任务定义中，我们可以这样配置：

{
  "name": "mysql-storage",
  "host": null,
  "dockerVolumeConfiguration": {
    "scope": "task",
    "autoprovision": true,
    "driver": "local"
  }
}

对应的容器挂载点配置：

{
  "sourceVolume": "mysql-storage",
  "containerPath": "/var/lib/mysql"
}

Grafana存储卷配置

Grafana需要持久化仪表板配置和用户数据：

{
  "name": "grafana-storage",
  "host": null,
  "dockerVolumeConfiguration": {
    "scope": "task",
    "autoprovision": true,
    "driver": "local"
  }
}

挂载点配置：

{
  "sourceVolume": "grafana-storage",
  "containerPath": "/var/lib/grafana"
}

DevLake日志卷配置

DevLake应用日志需要持久化存储：

{
  "name": "devlake-log",
  "host": null,
  "dockerVolumeConfiguration": {
    "scope": "task",
    "autoprovision": true,
    "driver": "local"
  }
}

挂载点配置：

{
  "sourceVolume": "devlake-log",
  "containerPath": "/app/logs"
}

容器端口映射调整

在ECS Fargate使用awsvpc网络模式时，有一个重要限制：主机端口和容器端口在端口映射中必须相同。这意味着我们需要调整Grafana的端口配置。

原始Docker Compose配置：

ports:
  - 3002:3000

调整为ECS兼容配置：

"portMappings": [
  {
    "containerPort": 3000,
    "hostPort": 3000,
    "protocol": "tcp"
  }
]

完整ECS任务定义示例

以下是整合了所有组件的完整ECS任务定义示例：

{
  "family": "devlake-task",
  "networkMode": "awsvpc",
  "requiresCompatibilities": ["FARGATE"],
  "cpu": "2048",
  "memory": "4096",
  "volumes": [
    {
      "name": "mysql-storage",
      "host": null,
      "dockerVolumeConfiguration": {
        "scope": "task",
        "autoprovision": true,
        "driver": "local"
      }
    },
    {
      "name": "grafana-storage",
      "host": null,
      "dockerVolumeConfiguration": {
        "scope": "task",
        "autoprovision": true,
        "driver": "local"
      }
    },
    {
      "name": "devlake-log",
      "host": null,
      "dockerVolumeConfiguration": {
        "scope": "task",
        "autoprovision": true,
        "driver": "local"
      }
    }
  ],
  "containerDefinitions": [
    {
      "name": "mysql",
      "image": "mysql:8.0.26",
      "essential": true,
      "portMappings": [
        {
          "containerPort": 3306,
          "hostPort": 3306,
          "protocol": "tcp"
        }
      ],
      "environment": [
        { "name": "MYSQL_ROOT_PASSWORD", "value": "admin" },
        { "name": "MYSQL_DATABASE", "value": "lake" },
        { "name": "MYSQL_USER", "value": "merico" },
        { "name": "MYSQL_PASSWORD", "value": "merico" }
      ],
      "mountPoints": [
        {
          "sourceVolume": "mysql-storage",
          "containerPath": "/var/lib/mysql"
        }
      ]
    },
    {
      "name": "grafana",
      "image": "apache/devlake-dashboard:v1.0.0",
      "essential": true,
      "portMappings": [
        {
          "containerPort": 3000,
          "hostPort": 3000,
          "protocol": "tcp"
        }
      ],
      "environment": [
        { "name": "GF_SERVER_ROOT_URL", "value": "http://localhost:4000/grafana" },
        { "name": "GF_USERS_DEFAULT_THEME", "value": "light" },
        { "name": "MYSQL_URL", "value": "mysql:3306" },
        { "name": "MYSQL_DATABASE", "value": "lake" },
        { "name": "MYSQL_USER", "value": "merico" },
        { "name": "MYSQL_PASSWORD", "value": "merico" },
        { "name": "TZ", "value": "UTC" }
      ],
      "mountPoints": [
        {
          "sourceVolume": "grafana-storage",
          "containerPath": "/var/lib/grafana"
        }
      ]
    },
    {
      "name": "devlake",
      "image": "apache/devlake:v1.0.0",
      "essential": true,
      "portMappings": [
        {
          "containerPort": 8080,
          "hostPort": 8080,
          "protocol": "tcp"
        }
      ],
      "environment": [
        { "name": "LOGGING_DIR", "value": "/app/logs" },
        { "name": "TZ", "value": "UTC" }
      ],
      "mountPoints": [
        {
          "sourceVolume": "devlake-log",
          "containerPath": "/app/logs"
        }
      ]
    },
    {
      "name": "config-ui",
      "image": "apache/devlake-config-ui:v1.0.0",
      "essential": true,
      "portMappings": [
        {
          "containerPort": 4000,
          "hostPort": 4000,
          "protocol": "tcp"
        }
      ],
      "environment": [
        { "name": "DEVLAKE_ENDPOINT", "value": "devlake:8080" },
        { "name": "GRAFANA_ENDPOINT", "value": "grafana:3000" },
        { "name": "TZ", "value": "UTC" }
      ],
      "dependsOn": [
        {
          "containerName": "devlake",
          "condition": "START"
        }
      ]
    }
  ]
}

部署建议与最佳实践

1. 资源分配：根据实际使用情况调整CPU和内存配置。生产环境建议至少4GB内存。

2. 安全考虑：

   • 使用AWS Secrets Manager管理数据库密码等敏感信息
   • 配置适当的安全组规则，限制对数据库端口的访问

3. 高可用性：

   • 考虑在多个可用区部署服务
   • 配置自动扩展策略以应对负载变化

4. 监控：

   • 启用ECS容器洞察功能
   • 配置CloudWatch警报监控关键指标

5. 备份策略：

   • 对EFS卷(如果使用)设置定期备份
   • 考虑使用RDS替代容器化的MySQL以获得更好的可靠性和管理性

通过以上配置，您可以在AWS ECS Fargate上成功部署Apache DevLake，获得一个完全托管、可扩展的数据湖解决方案。这种部署方式相比传统的Docker Compose部署提供了更好的可靠性、可扩展性和管理便利性。