如何通过google-compute-engine Python客户端过滤Google Cloud实例？

依赖版本确认 #

你使用的依赖版本如下，方案兼容该版本范围：

Package                  Version
------------------------ ---------
google-api-core          2.0.1
google-auth              2.0.2
google-cloud-compute     0.5.0
google-compute-engine    2.8.13

方案1：服务端过滤（推荐） #

aggregated_list接口原生支持通过filter参数在服务端完成标签过滤，减少无用数据传输，过滤语法规则如下：

• 匹配包含指定单个标签的实例：filter='tags.items="target-tag"'
• 匹配同时包含多个标签的实例：filter='tags.items="tag1" AND tags.items="tag2"'
• 匹配包含任意一个指定标签的实例：filter='tags.items="tag1" OR tags.items="tag2"'
  示例代码：

from google.cloud import compute_v1

def list_instances_by_tag(project_id: str, target_tag: str):
    client = compute_v1.InstancesClient()
    # 传入服务端过滤参数
    request = compute_v1.AggregatedListInstancesRequest(
        project=project_id,
        filter=f'tags.items="{target_tag}"'
    )

    for zone, response in client.aggregated_list(request=request):
        if response.instances:
            for instance in response.instances:
                print(f"匹配实例：{zone} - {instance.name}")

方案2：本地过滤（兜底方案） #

如果过滤规则非常复杂，服务端语法无法满足，可以先拉取全量实例数据后本地做过滤，示例代码：

from google.cloud import compute_v1

def list_instances_by_tag_local(project_id: str, target_tags: list):
    client = compute_v1.InstancesClient()
    request = compute_v1.AggregatedListInstancesRequest(project=project_id)
    matched_instances = []

    for zone, response in client.aggregated_list(request=request):
        if response.instances:
            for instance in response.instances:
                # 取实例标签做交集判断
                instance_tags = instance.tags.items if instance.tags else []
                if set(target_tags).issubset(set(instance_tags)):
                    matched_instances.append({
                        "zone": zone,
                        "name": instance.name,
                        "tags": instance_tags
                    })
    return matched_instances

注意事项 #

• 服务端过滤的标签值区分大小写
• 若实例未配置任何标签，tags字段会返回空，过滤时需做判空处理避免报错

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