Cloud Service Mesh 提供服务的健康状况和性能的可观测性。为了获得遥测数据，Cloud Service Mesh 依赖于边车代理，您可将此类代理作为单独容器注入到与工作负载相同的 Pod 中。代理会拦截所有工作负载的入站和出站 HTTP 流量，并将数据报告给 Cloud Service Mesh。借助此系统，服务开发者不必对其代码进行插桩处理即可收集遥测数据。
安装 Cloud Service Mesh 时，系统会在您的Google Cloud 项目中启用 Cloud Monitoring 和 Cloud Logging。为了报告遥测数据，注入到服务 Pod 的每个边车代理都会调用 Cloud Monitoring API 和 Cloud Logging API。遥测数据会自动上传至Google Cloud 控制台中的 Cloud Service Mesh 页面。请注意， Google Cloud 控制台中的 Cloud Service Mesh 页面仅显示 HTTP 服务的指标。
Cloud Service Mesh 在Google Cloud 控制台中提供多个预配置的服务信息中心，您无需手动设置信息中心和图表。这些详细遥测让运营商可观察服务行为，以及维护和优化其应用并进行问题排查。
[[["易于理解","easyToUnderstand","thumb-up"],["解决了我的问题","solvedMyProblem","thumb-up"],["其他","otherUp","thumb-up"]],[["很难理解","hardToUnderstand","thumb-down"],["信息或示例代码不正确","incorrectInformationOrSampleCode","thumb-down"],["没有我需要的信息/示例","missingTheInformationSamplesINeed","thumb-down"],["翻译问题","translationIssue","thumb-down"],["其他","otherDown","thumb-down"]],["最后更新时间 (UTC)：2025-08-19。"],[],[],null,["# Observability overview\n======================\n\nCloud Service Mesh provides *observability* into the health and performance\nof your services. To obtain telemetry data, Cloud Service Mesh relies on\n[sidecar proxies](/service-mesh/docs/onboarding/kubernetes-workloads#inject_sidecar_proxies)\nthat you inject as a separate container into the same Pods as your workloads.\nThe proxies intercept all inbound and outbound HTTP traffic to the workloads and\nreport the data to Cloud Service Mesh. With this system, service developers\ndon't have to instrument their code to collect telemetry data.\n\nCloud Monitoring and Cloud Logging are enabled in your\nGoogle Cloud project when you install Cloud Service Mesh. To report\ntelemetry data, each sidecar proxy that is injected into your service Pods calls\nthe Cloud Monitoring API and the Cloud Logging API. The telemetry data is\nautomatically uploaded to the Cloud Service Mesh pages in the\nGoogle Cloud console. Note that metrics are displayed only for HTTP services\non the Cloud Service Mesh pages in the Google Cloud console.\n\nCloud Service Mesh provides several preconfigured service dashboards in the\nGoogle Cloud console so you don't have to manually set up dashboards and\ncharts. This detailed telemetry enables operators to observe service behavior,\nand empowers them to troubleshoot, maintain, and optimize their applications.\n\nOn the Cloud Service Mesh pages in the Google Cloud console, you can:\n\n- Get an overview of all services in your mesh, providing you critical,\n service-level metrics on three of the\n [four golden signals of monitoring](https://sre.google/sre-book/monitoring-distributed-systems/#xref_monitoring_golden-signals):\n latency, traffic, and errors.\n\n- Define, review, and set alerts against service level objectives (SLOs),\n which summarize your service's user-visible performance.\n\n- View metric charts for individual services and deeply analyze them with\n filtering and breakdowns, including by response code, protocol, destination\n pod, traffic source, and more.\n\n- Get detailed information about the endpoints for each service, and see\n how traffic is flowing between services, and what performance looks like\n for each communication edge.\n\n- Explore a service topology graph visualization that shows your mesh's\n services and their relationships.\n\nWhat's next\n-----------\n\n- [Control access to Cloud Service Mesh in the Google Cloud console](/service-mesh/docs/access-dashboard)\n- [Learn about SLOs](/service-mesh/docs/observability/slo-overview)\n- [Create an alerting policy on an SLO](/service-mesh/docs/observability/alert-policy-slo)"]]
