简体   繁体   English

除了 /stats/summary 之外,还有其他方法可以获得节点的“fs”指标吗

[英]Is there another way to get "fs" metric of node besides /stats/summary

We've been using "/stats/summary" to get "fs" metrics, which is like:我们一直在使用“/stats/summary”来获取“fs”指标,例如:

"fs": {
 "time": "2021-10-14T03:46:05Z",
 "availableBytes": 17989276262,
 "capacityBytes": 29845807308,
 "usedBytes": 5856531046,
 "inodesFree": 16799593,
 "inodes": 17347097,
 "inodesUsed": 57504
},

And due to this Move away from kubelet stats/summary , we need to get the same data in another way.并且由于这个远离 kubelet stats/summary ,我们需要以另一种方式获取相同的数据。

We've tried /metrics/cadvisor and /metrics/resources, but were not successful to get "fs" data.我们已经尝试过 /metrics/cadvisor 和 /metrics/resources,但是没有成功获得“fs”数据。 Also, it seems that CAdvisor will also be deprecated (in TBD+2 here )此外,似乎 CAdvisor 也将被弃用(在 TBD+2 here 中

We've been searching the net for possible solution but can't seem to find any.我们一直在网上寻找可能的解决方案,但似乎找不到任何解决方案。

Any ideas on how this can be done?关于如何做到这一点的任何想法? Or probably point us to the right direction or documentation?或者可能为我们指出正确的方向或文档?

Thank you in advance.先感谢您。

Posted community wiki based on Github topic.基于 Github 主题发布的社区 wiki。 Feel free to expand it.随意扩展它。


Personally, I have not found any equivalent of this call ( /api/v1/nodes/<node name>/proxy/stats/summary ), and as it is still working and not deprecated in the Kubernetes newest versions ( 1.21 and 1.22 ), I'd recommend just using it and wait for information about replacement from the Kubernetes team.就我个人而言,我还没有找到任何与此调用等效的方法 ( /api/v1/nodes/<node name>/proxy/stats/summary ),因为它仍在工作,并且在 Kubernetes 最新版本 ( 1.211.22 ) 中并未弃用,我建议只使用它并等待 Kubernetes 团队提供有关更换的信息。 Check below information:检查以下信息:

Information from this GitHub topic - # Reduce the set of metrics exposed by the kubelet #68522 (last edited: November 2020, issue open):来自此GitHub 主题的信息 - # 减少 kubelet 公开的指标集 #68522 (上次编辑:2020 年 11 月,问题公开):

It seems that /stats/summary/ does not have any replacement recommendation ready:似乎/stats/summary/没有准备好任何替换建议:

[TBD] Propose out-of-tree replacements for kubelet monitoring endpoints [TBD] 提出 kubelet 监控端点的树外替换

They will keep the Summary API for the next four versions counting from the version in which replacement will be implemented:他们将保留接下来四个版本的摘要 API,从将实施替换的版本开始计算:

[TBD+4] Remove the Summary API, cAdvisor prometheus metrics and remove the --enable-container-monitoring-endpoints flag. [TBD+4] 删除摘要 API、cAdvisor prometheus 指标并删除--enable-container-monitoring-endpoints标志。


In Kubernetes v1.23 changelog there is no information about changing anything related to the Summary API.在 Kubernetes v1.23更改日志中,没有关于更改与摘要 API 相关的任何内容的信息。

I'd suggest observing and pinging Kubernetes developers directly in this GitHub topic for more information.我建议直接在此 GitHub 主题中观察和 ping Kubernetes 开发人员以获取更多信息。

声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.

相关问题 kubernetes UI无法获取节点统计信息(503) - kubernetes UI fails to get node stats (503) Kubelet API / stats / summary /用法NanoCores不正确 - Kubelet API /stats/summary/ usageNanoCores not correct Minikube 的 /stats/summary 中缺少 Kubelet Rootfs.UsedBytes - Kubelet Rootfs.UsedBytes missing in /stats/summary in Minikube HPA在GKE节点自动缩放期间无法获取CPU指标 - HPA could not get CPU metric during GKE node auto-scaling 有没有办法在 pod 迁移到另一个应用程序节点之前告诉它在哪里 - Is there a way to tell where a pod was before it is migrated to another app node 如何在kubernetes中获取kubernetes组件指标数据? - How to get kubernetes components metric data in kubernetes? 如何获得 kubernetes 集群范围的指标 - How to get kubernetes cluster wide metric Kube.netes pod:如何获取 pod 统计信息? - Kubernetes pods: how to get pod stats? HorizontalPodAutoscaler 基于自定义指标的缩放 - 节点池级别指标 - HorizontalPodAutoscaler scaling based on custom metrics - node-pool level metric 如何根据 Kubernetes kublet 摘要端点返回的指标计算 CPU 使用率、内存使用率,即,<machine-ip> :10255/统计/摘要 - How to calculate cpu usage, memory usage from the metrics returned by Kubernetes kublet summary endpoint ie., <machine-ip>:10255/stats/summary
 
粤ICP备18138465号  © 2020-2024 STACKOOM.COM