Longhorn Upgrade Checker 信息采集机制：`Allow Collecting Longhorn Usage Metrics` 设置与 extraInfo 指标格式全解析

云原生存储高可用容器编排【免费下载链接】longhornCloud-Native distributed storage built on and for Kubernetes项目地址https://gitcode.com/gh_mirrors/lo/longhorn点击查看免费下载Longhorn 的 Upgrade Checker升级检查器除了在发现新版本时于 UI 弹出升级提示外还承担着向官方升级应答服务器上报匿名使用数据、支撑 Longhorn 公共指标看板metrics.longhorn.io的职责。本文以仓库中的增强提案 20230420-upgrade-checker-info-collection.md 为核心系统讲解新增的Allow Collecting Longhorn Usage Metrics设置、extraInfo扩展后采集的节点与集群指标清单、设置采集的排除与转换规则、完整的数据格式示例并结合 Helm Chart 配置说明如何启用、覆盖或关闭该采集能力。背景与动机升级检查背后的数据价值Longhorn 官方维护的公共指标看板metrics.longhorn.io向公众展示 Longhorn 被实际使用的概况这类信息对两类人群有直接价值新用户与计划升级的用户可以据此了解主流用户群体采用的 Longhorn 与 Kubernetes 版本组合为自身的版本选型与升级路径提供参考Longhorn 开发团队借此了解真实世界的使用模式为功能改进与发展方向决策提供数据支撑。为了让使用模式的分析更加深入增强提案提出在升级检查时收集更多关于**卷volumes、宿主机系统host systems与功能特性features**的信息并明确了两个目标Goals在升级检查期间扩展对用户集群信息的采集范围提供一个新的设置项让用户可以选择启用或关闭该采集能力。该提案不涉及任何 API 变更API changes 为None升级策略Upgrade strategy同样为None属于完全向后兼容的能力增强。总体方案新设置 extraInfo扩展整个方案由两个相互配合的部分组成扩展extraInfo采集内容在原有 Upgrade Checker 上报的extraInfo基础上新增节点信息Node Info与集群信息Cluster Info两类指标并通过 upgrade responder 请求发送至服务端新增Allow Collecting Longhorn Usage Metrics设置作为采集的总开关。若该值被设置为false则不再采集任何额外信息extra information。用户既可以在 Longhorn 公共指标看板上查看 Longhorn 的整体使用情况也可以随时关闭本集群的数据采集确保隐私可控。Allow Collecting Longhorn Usage Metrics设置详解设置定义该设置在提案中给出的完整定义如下属性值DisplayName显示名Allow Collecting Longhorn Usage MetricsDescription描述启用该设置后Longhorn 将向 metrics.longhorn.io 提供额外的使用指标。这些信息有助于我们更好地了解 Longhorn 的使用方式从而为未来的改进做出贡献。Category分类SettingCategoryGeneral常规设置Type类型SettingTypeBool布尔类型Required是否必填trueReadOnly是否只读falseDefault默认值true从定义可以看出该设置默认启用即默认情况下 Longhorn 会采集并上报上述额外信息同时它是布尔类型且可写用户可以自由地将其改为false来完全关闭额外信息采集。在 Helm Chart 中的体现与覆盖方式在仓库的 Helm Chart 中该设置通过defaultSettings.allowCollectingLonghornUsageMetrics暴露位于 chart/values.yaml其注释说明了用途Setting that allows Longhorn to periodically collect anonymous usage data for product improvement purposes. Longhorn sends collected data to the Upgrade Responder server, which is the data source of the Longhorn Public Metrics Dashboard (metrics.longhorn.io). The Upgrade Responder server does not store data that can be used to identify clients, including IP addresses.这里补充了三个关键事实采集是**周期性periodically**进行的匿名使用数据上报数据发送至Upgrade Responder服务器它是 Longhorn 公共指标看板的数据源服务器不会存储可用于识别客户端身份的数据包括 IP 地址从机制上保证了匿名性。Chart 中该值的默认状态为~即未显式指定表示由 Longhorn Manager 侧的内置默认值true生效。当用户在 Helm values 中显式赋值后chart/templates/default-setting.yaml 会通过kindIs invalid判断仅在值为非空时才向默认设置 ConfigMap 注入对应键{{- if not (kindIs invalid .Values.defaultSettings.allowCollectingLonghornUsageMetrics) }} allow-collecting-longhorn-usage-metrics: {{ .Values.defaultSettings.allowCollectingLonghornUsageMetrics }}同时 chart/questions.yaml 提供了对应的 UI 表单变量使用户在通过 Rancher 等图形界面部署 Longhorn 时也能直接配置该选项。需要说明的是该设置在早期 Helm Chart 中一度缺失仓库的 CHANGELOG-1.6.0.md 与 CHANGELOG-1.5.4.md 分别记录了 allow-collecting-longhorn-usage-metricssetting is missing from chart settings 的修复即后续版本已将该设置补入 Chart安装时即可通过 values 控制。通过 UI 或 kubectl 配置与 Longhorn 其他设置一样该设置也可以在运行时通过两种方式调整Longhorn UI进入「Setting」页面找到Allow Collecting Longhorn Usage Metrics项切换开关后保存即可kubectlLonghorn 的所有设置都以settings.longhorn.io自定义资源存在可以推断使用以下命令编辑kubectl -n longhorn-system edit settings.longhorn.io allow-collecting-longhorn-usage-metrics将value字段改为false并保存后采集即被关闭之后 Upgrade Checker 上报时将不再携带额外信息。extraInfo采集内容详解提案将采集内容划分为两大类别节点信息与集群信息。其中节点信息由每个集群节点分别上报集群信息则由集群中的某一个节点汇总上报。节点信息Node Info以下信息从每个集群节点发送指标键含义longhorn_node_disk_hdd/ssd/nvme/unknown_count该节点上不同类型设备的磁盘数量HDD / SSD / NVMe / 未知host_kernel_release宿主机内核版本host_os_distro宿主机操作系统发行版kubernetes_node_providerKubernetes 节点提供商其中磁盘类型计数存在一个需要特别注意的边界提案明确备注该值在集群节点为虚拟机时可能不准确this value may not be accurate if the cluster node is a virtual machine。这是因为磁盘类型判定依赖设备的底层特性虚拟化环境下设备类型往往被抽象或透传统计结果可能偏离物理真实值解读数据时应结合节点形态综合判断。集群信息Cluster Info以下信息从集群中的某一个节点发送即全局汇总数据指标键含义longhorn_namespace_uidLonghorn 命名空间的 UID用于统计采纳率/适配率 adaptation ratelonghorn_node_count集群节点数量longhorn_volume_access_mode_rwo/rwx/unknown_count不同访问模式RWO / RWX / 未知的卷数量longhorn_volume_data_locality_disabled/best_effort/strict_local/unknown_count不同数据本地性策略的卷数量longhorn_volume_frontend_blockdev/iscsi_count不同前端blockdev / iscsi的卷数量longhorn_volume_average_size卷的平均大小longhorn_volume_average_actual_size卷的平均实际占用大小longhorn_volume_average_snapshot_count每个卷的平均快照数量longhorn_volume_average_number_of_replicas每个卷的平均副本数量longhorn_engine_image/instance_manager/manager/ui_average_cpu_usage_coreLonghorn 各组件engine-image / instance-manager / manager / ui的平均 CPU 使用量核longhorn_engine_image/instance_manager/manager/ui_average_memory_usage_mibLonghorn 各组件engine-image / instance-manager / manager / ui的平均内存使用量MiBlonghorn_setting_name各 Longhorn 设置项见下述采集规则设置的采集规则设置类指标longhorn_setting_name并非全部照搬而是分为三类处理1. 排除清单Settings to exclude以下设置不采集多数涉及镜像地址、凭据、路径、版本信息等敏感或冗余内容排除的设置SettingNameBackupTargetCredentialSecret备份目标凭据 SecretSettingNameDefaultEngineImage默认引擎镜像SettingNameDefaultInstanceManagerImage默认实例管理器镜像SettingNameDefaultShareManagerImage默认共享管理器镜像SettingNameDefaultBackingImageManagerImage默认 Backing Image 管理器镜像SettingNameSupportBundleManagerImageSupport Bundle 管理器镜像SettingNameCurrentLonghornVersion当前 Longhorn 版本SettingNameLatestLonghornVersion最新 Longhorn 版本SettingNameStableLonghornVersions稳定 Longhorn 版本列表SettingNameDefaultLonghornStaticStorageClass默认静态 StorageClassSettingNameDeletingConfirmationFlag删除确认标志SettingNameDefaultDataPath默认数据路径SettingNameUpgradeCheckerUpgrade Checker 开关SettingNameAllowCollectingLonghornUsage本采集开关自身SettingNameDisableReplicaRebuild已废弃SettingNameGuaranteedEngineCPU已废弃排除BackupTargetCredentialSecret显然是出于安全考虑——它包含备份目标凭据绝不能外泄排除默认镜像/版本类设置则是因为它们可从其他渠道获得且变化频繁。2. 需要转换处理的设置个别设置需要加工后才能体现其一般用途典型代表是SettingNameBackupTarget——采集时转换为备份目标的类型/协议如cifs、nfs、s3而不是原始 URL 或完整配置。3. 布尔化采集的设置以下设置以布尔值采集true表示已配置false表示未配置避免暴露具体配置内容设置布尔化语义SettingNameTaintToleration是否配置了污点容忍SettingNameSystemManagedComponentsNodeSelector是否配置了系统组件节点选择器SettingNameRegistrySecret是否配置了镜像仓库 SecretSettingNamePriorityClass是否配置了优先级类SettingNameStorageNetwork是否配置了存储网络其余设置则按原样as it is采集。这套规则的设计意图非常清晰排除敏感与冗余信息收敛高敏配置为布尔标记兼顾统计价值与用户隐私。完整数据示例upgrade_request 记录解析提案给出了启用采集后 upgrade responder 数据库中upgrade_request表的实际记录示例节选自 enhancements/20230420-upgrade-checker-info-collection.mdname: upgrade_request time app_version host_kernel_release host_os_distro kubernetes_node_provider kubernetes_version longhorn_engine_image_average_cpu_usage_core longhorn_engine_image_average_memory_usage_mib longhorn_instance_manager_average_cpu_usage_core longhorn_instance_manager_average_memory_usage_mib longhorn_manager_average_cpu_usage_core longhorn_manager_average_memory_usage_mib longhorn_namespace_uid longhorn_node_count longhorn_node_disk_nvme_count longhorn_setting_allow_node_drain_with_last_healthy_replica longhorn_setting_allow_recurring_job_while_volume_detached longhorn_setting_allow_volume_creation_with_degraded_availability longhorn_setting_auto_cleanup_system_generated_snapshot longhorn_setting_auto_delete_pod_when_volume_detached_unexpectedly longhorn_setting_auto_salvage longhorn_setting_backing_image_cleanup_wait_interval longhorn_setting_backing_image_recovery_wait_interval longhorn_setting_backup_compression_method longhorn_setting_backup_concurrent_limit longhorn_setting_backup_target longhorn_setting_backupstore_poll_interval longhorn_setting_concurrent_automatic_engine_upgrade_per_node_limit longhorn_setting_concurrent_replica_rebuild_per_node_limit longhorn_setting_concurrent_volume_backup_restore_per_node_limit longhorn_setting_crd_api_version longhorn_setting_create_default_disk_labeled_nodes longhorn_setting_default_data_locality longhorn_setting_default_replica_count longhorn_setting_disable_revision_counter longhorn_setting_disable_scheduling_on_cordoned_node longhorn_setting_engine_replica_timeout longhorn_setting_failed_backup_ttl longhorn_setting_fast_replica_rebuild_enabled longhorn_setting_guaranteed_engine_manager_cpu longhorn_setting_guaranteed_instance_manager_cpu longhorn_setting_guaranteed_replica_manager_cpu longhorn_setting_kubernetes_cluster_autoscaler_enabled longhorn_setting_node_down_pod_deletion_policy longhorn_setting_node_drain_policy longhorn_setting_orphan_auto_deletion longhorn_setting_priority_class longhorn_setting_recurring_failed_jobs_history_limit longhorn_setting_recurring_successful_jobs_history_limit longhorn_setting_registry_secret longhorn_setting_remove_snapshots_during_filesystem_trim longhorn_setting_replica_auto_balance longhorn_setting_replica_file_sync_http_client_timeout longhorn_setting_replica_replenishment_wait_interval longhorn_setting_replica_soft_anti_affinity longhorn_setting_replica_zone_soft_anti_affinity longhorn_setting_restore_concurrent_limit longhorn_setting_restore_volume_recurring_jobs longhorn_setting_snapshot_data_integrity longhorn_setting_snapshot_data_integrity_cronjob longhorn_setting_snapshot_data_integrity_immediate_check_after_snapshot_creation longhorn_setting_storage_minimal_available_percentage longhorn_setting_storage_network longhorn_setting_storage_over_provisioning_percentage longhorn_setting_storage_reserved_percentage_for_default_disk longhorn_setting_support_bundle_failed_history_limit longhorn_setting_system_managed_components_node_selector longhorn_setting_system_managed_pods_image_pull_policy longhorn_setting_taint_toleration longhorn_ui_average_cpu_usage_core longhorn_ui_average_memory_usage_mib longhorn_volume_access_mode_rwo_count longhorn_volume_average_actual_size longhorn_volume_average_number_of_replicas longhorn_volume_average_size longhorn_volume_average_snapshot_count longhorn_volume_data_locality_disabled_count longhorn_volume_frontend_blockdev_count value ---- ----------- ------------------- -------------- ------------------------ ------------------ -------------------------------------------- ---------------------------------------------- ------------------------------------------------ -------------------------------------------------- --------------------------------------- ----------------------------------------- ---------------------- ------------------- ----------------------------- ----------------------------------------------------------- ---------------------------------------------------------- ----------------------------------------------------------------- ------------------------------------------------------- ------------------------------------------------------------------ ----------------------------- ---------------------------------------------------- ----------------------------------------------------- ------------------------------------------ ---------------------------------------- ------------------------------ ------------------------------------------ ------------------------------------------------------------------- ---------------------------------------------------------- ---------------------------------------------------------------- -------------------------------- -------------------------------------------------- -------------------------------------- -------------------------------------- ----------------------------------------- ---------------------------------------------------- --------------------------------------- ---------------------------------- --------------------------------------------- ---------------------------------------------- ------------------------------------------------ ----------------------------------------------- ------------------------------------------------------ ---------------------------------------------- ---------------------------------- ------------------------------------- ------------------------------- ---------------------------------------------------- -------------------------------------------------------- -------------------------------- -------------------------------------------------------- ------------------------------------- ------------------------------------------------------ ---------------------------------------------------- ------------------------------------------- ------------------------------------------------ ----------------------------------------- ---------------------------------------------- ---------------------------------------- ------------------------------------------------ -------------------------------------------------------------------------------- ----------------------------------------------------- -------------------------------- ----------------------------------------------------- ------------------------------------------------------------- ---------------------------------------------------- -------------------------------------------------------- ------------------------------------------------------ --------------------------------- ---------------------------------- ------------------------------------ ------------------------------------- ----------------------------------- ------------------------------------------ ---------------------------- -------------------------------------- -------------------------------------------- --------------------------------------- ----- 1683598256887331729 v1.5.0-dev 5.3.18-59.37-default sles k3s v1.23.15k3s1 5m 11 4m 83 22m 85 1b96b299-b785-468b-ab80-b5b5b12fbe00 3 1 false false true true true true 60 300 lz4 5 none 300 0 5 5 longhorn.io/v1beta2 false disabled 3 false true 8 1440 true 12 12 12 false do-nothing block-if-contains-last-replica false false 1 1 false false disabled 30 600 false true 5 false fast-check 0 0 */7 * * false 25 false 200 30 1 false if-not-present false 0 4 3 79816021 2 8589934592 0 3 3 1 1683598257082240493 v1.5.0-dev 5.3.18-59.37-default sles k3s v1.23.15k3s1 1 1 1683598257825718008 v1.5.0-dev 5.3.18-59.37-default sles k3s v1.23.15k3s1 1 1对该示例数据的解读要点如下时间戳与版本每条记录以毫秒级时间戳为主键app_version为 Longhorn 版本示例为v1.5.0-dev开发版节点信息字段host_kernel_release5.3.18-59.37-default、host_os_distrosles、kubernetes_node_providerk3s与前述 Node Info 清单一一对应组件资源使用longhorn_manager_average_cpu_usage_core22m约 0.022 核、longhorn_ui_average_memory_usage_mib85约 85 MiB等展示各组件平均资源占用的量纲卷维度统计longhorn_volume_access_mode_rwo_count3、longhorn_volume_average_size85899345928 GiB 字节数、longhorn_volume_average_number_of_replicas3等设置字段longhorn_setting_backup_targetnone对应无备份目标、longhorn_setting_backup_compression_methodlz4、longhorn_setting_snapshot_data_integrityfast-check、longhorn_setting_crd_api_versionlonghorn.io/v1beta2等注意backup_target已按规则转换为协议/类型而非完整 URL关键隐私设计后两条记录同一集群的后续上报仅保留了少量字段与longhorn_namespace_uid且该 UID 为随机 UUID1b96b299-b785-468b-ab80-b5b5b12fbe00用于跨记录关联同一集群而不暴露可识别身份的信息。数据流与 Upgrade Responder 部署采集到的数据由 Longhorn Manager 的 Upgrade Checker 周期性发送至 Upgrade Responder 服务器默认端点https://longhorn-upgrade-responder.rancher.io/v1/checkupgrade可通过defaultSettings.upgradeResponderURL覆盖见 chart/values.yaml。服务端将请求写入 InfluxDB 数据库由 Grafana 渲染为 Longhorn 公共指标看板。若需自建 Upgrade Responder 服务端进行验证仓库提供了两条路径手动部署参考 dev/upgrade-responder/README.md在安装 Longhorn 后执行./install.sh脚本会依次部署 InfluxDB数据库longhorn_upgrade_responder、Grafana 以及 Upgrade Responder Helm Releaselonghorn-upgrade-responder并输出 Upgrade Checker 使用的服务端 URL[Upgrade Checker] URL : http://longhorn-upgrade-responder.default.svc.cluster.local:8314/v1/checkupgrade [InfluxDB] URL : http://influxdb.default.svc.cluster.local:8086 Database : longhorn_upgrade_responder Username : root Password : root [Grafana] Dashboard : http://1.2.3.4:30864 Username : admin Password : admin部署完成后可通过kubectl port-forward将 8314 端口映射到本机kubectl port-forward $POD_NAME 8080:8314即可获得本地可达的升级检查端点。官方 Chart 配置deploy/upgrade_responder_server/chart-values.yaml 提供了生产化配置样例其中applicationName: longhorn会创建名为longhorn_upgrade_responder的数据库responseConfig.configMap维护各版本的稳定标记stable tags供升级提示与版本统计使用同时支持通过flags.scarfEndpoint配置下载统计网关。测试计划与验收标准提案给出的验证路径Test plan简洁明确部署 Upgrade Responder 服务端可选用仓库提供的自建方案分别在Allow Collecting Longhorn Usage Metrics启用与禁用两种状态下核对数据库中upgrade_request表的记录启用时应能看到本文所述的节点信息、集群信息与设置字段可对照上文示例格式检查列结构与取值禁用时extraInfo不再携带额外信息即验证开关真正生效。总结Allow Collecting Longhorn Usage Metrics设置与extraInfo扩展共同构成了 Longhorn 匿名使用数据采集的完整机制默认启用、周期上报、内容分节点与集群两类、设置采集按排除 / 转换 / 布尔化三级规则收敛、服务器不存 IP 等可识别数据。对于运维人员理解这一机制有助于在隐私合规要求下通过 UI、kubectl 或 Helm values 快速关闭采集对于开发与研究人员则可以通过本文的字段清单与示例数据格式对 Longhorn 公共指标看板的数据做深入的解读与分析。赞分享云原生存储高可用容器编排【免费下载链接】longhornCloud-Native distributed storage built on and for Kubernetes项目地址https://gitcode.com/gh_mirrors/lo/longhorn点击查看免费下载相关推荐WebSite-Downloader如何将整个网站变成你的永久数字资产WebSite Downloader如何将整个网站变成你的永久数字资产 在信息瞬息万变的互联网时代你是否曾遇到过心爱的网站突然消失重要的技术文档被下架云原生存储高可用容器编排Longhorn 卷引擎镜像自动升级Engine Image Automatic Upgrade机制解析Longhorn 卷引擎镜像自动升级Engine Image Automatic Upgrade机制解析 导读 本文以 Longhorn 的增强提案 enh云原生存储高可用容器编排Longhorn Engine Upgrade Enforcement 解析升级前的引擎镜像版本强制校验机制Longhorn Engine Upgrade Enforcement 解析升级前的引擎镜像版本强制校验机制 本文围绕 Longhorn 的 Engine U云原生存储高可用容器编排上一篇MCP Toolbox 数据库工具实战mongodb-update-many 批量更新工具完全指南下一篇mold 项目内嵌 oneTBB 的 blocked_range2d 详解二维可递归划分区间的原理、分裂算法与并行实践创作声明：本文部分内容由AI辅助生成（AIGC），仅供参考