写在前面
学习K8s，所以整理分享给小伙伴
这里要说明的是：这一套方案太吃硬件了，需要高配的本才能跑起来
我的16G运存,集群用三个虚机部署的，工作节点都是3核5G的配置
折腾了两天没有跑起来，后来放弃了，查了下，还是资源的问题。浪费了两天的假期:(
然后睡了一觉，起来好了，勉强能跑起来，但是巨卡
博文涉及主要内容k8s集群helm方式日志管理方案(Elasticsearch+Filebeat+Kibana+Metricbeat)搭建Demo一些搭建过程的避坑说明
部分内容參考《CKA/CKAD应试指南》《基于Kubernetes的容器云平台实战》《Kubernetes权威指南》
我所渴求的，無非是將心中脫穎語出的本性付諸生活，為何竟如此艱難呢 ------赫尔曼·黑塞《德米安》
方案简述
部署完Kubernetes平台后,会安装运行大量的应用, Kubernetes平台及各种应用系统运行过程中会产生大量的各种各样的系统日志和应用日志,通过对这些日志的采集、传输、存储、分析,掌握系统运行状况,对故障处理、安全审计、行为统计等有非常重要的作用。下面和小伙伴聊聊日志采集、传输、存储及分析的具体方案及实践。
Kubernetes 平台上的日志按照运维领域分为应用日志和平台日志.
容器引擎可以通过DockerDeamon来接收和重定向这些日志数据。
在容器中输出到控制台的日志，都会以*-json.log的命名方式保存在/var/lib/docker/containers/目录下
┌──[root@vms81.liruilongs.github.io]-[/]└─$cd /var/lib/docker/containers/┌──[root@vms81.liruilongs.github.io]-[/var/lib/docker/containers]└─$ls0606dd217f4f2f315f86fb378ffa65ed0c59ba7580869b9a88634dd6e171fdc0 5fde81847fefef953a4575a98e850ce59e1f12c870c9f3e0707a993b8fbecdf0 a72407dce4ac3cd294f8cd39e3fe9b882cbab98d97ffcfad5595f85fb15fec86。。。。。。。。。。。。。。。。。。。。a66a7eee79ced84e6d9201ee38f9ab887f7bfc0103e1172e107b337108a63638 ff545c3a08264a28968800fe0fb2bbd5f4381029f089e9098c9e1484d310fcc1┌──[root@vms81.liruilongs.github.io]-[/var/lib/docker/containers]└─$cd 0606dd217f4f2f315f86fb378ffa65ed0c59ba7580869b9a88634dd6e171fdc0;ls0606dd217f4f2f315f86fb378ffa65ed0c59ba7580869b9a88634dd6e171fdc0-json.log checkpoints config.v2.json hostconfig.json mounts┌──[root@vms81.liruilongs.github.io]-[/var/lib/docker/containers/0606dd217f4f2f315f86fb378ffa65ed0c59ba7580869b9a88634dd6e171fdc0]└─$
复制代码
对于应用日志这种数据量比较大的数据可以考虑通过网络直接重定向到收集器,不必通过文件中转,这样可以一定程度地加快数据流转的速度。对于容器应用数据的收集,由于容器的动态申请和释放的特性,因此日志中能够包含日志来源的拓扑信息和 IP 地址等身份信息就显得尤为重要。应用日志由于数据量巨大,般应当根据系统确定保存时间,并对存储空间溢出的异常进行保护。
在Kubernetes集群环境中，一个完整的应用或服务都会涉及为数众多的组件运行，各组件所在的 Node 及实例数量都是可变的。日志子系统如果不做集中化管理，则会系统的运维支撑造成很大的困难，因此有必要在集群层面对日志进行统一收集和检索等工作。所以就有了ELF和EFK
ELK
ELK 是ElasticSearch、 Logstash, Kibana的简称,是容器日志管理的核心套件。
ElasticSearch： 是基于 Apache Lucene 引擎开发的实时全文搜索和分析引擎,提供结构化数据的搜集、分析、存储数据三大功能。它提供 Rest 和 Java API 两种接口方式。
Logstash： 是一个日志搜集处理框架,也是日志透传和过滤的工具,它支持多种类型的日志,包括系统日志、错误日志和自定义应用程序日志。它可以从许多来源接收日志,这些来源包括 syslog、消息传递(例如 RabbitMQ)和 JMX,它能够以多种方式输出数据,包括电子邮件、Websockets 和 ElasticSearch
Kibana： 是一个图形化的 Web 应用,它通过调用 ElastieSearch 提供的 Rest 或 Java API 接口搜索存储在 ElasticSearch 中的日志,并围绕分析主题进行分析得出结果,并将这些结果数据按照需要进行界面可视化展示;另外, Kibana 可以定制仪表盘视图,展示不同的信息以便满足不同的需求。
EFK
EFK： Logstas 性能低，消耗资源，且存在不支持消息队列缓存及存在数据丢失的问题,所以 logstash 一般可以用fluentd或者filebeat替代
关于日志方案涉及的组件，这里列出官网地址，小伙伴可以去看看
elasticsearch：https://www.elastic.co/cn/elasticsearch/
metricbeat：https://www.elastic.co/cn/beats/metricbeat
kibana：https://www.elastic.co/cn/kibana/
filebeat：https://www.elastic.co/cn/beats/filebeat
在 K8s 集群部署统一的日志管理系统，需要以下两个前提条件。
API Server正确配置了CA证书。
DNS服务,跨主机组网启动、运行。
需要注意的是，这套方案涉及到 RBAC 权限处理的一些资源对象，要结合 helm 中 chart 的资源文件和当前集群的资源对象版本去修改。关于基于 RBAC 的权限处理Kubernetes在1.5版本中引入，在1.6版本时升级为Beta版本，在1.8版本时升级为GA，我们这里搭建用的7.9.1版本，RBAC用的Bate的版本，但是集群是1.22的版本，所以需要修改资源文件。
┌──[root@vms81.liruilongs.github.io]-[~/efk]└─$kubectl api-versions | grep rbacrbac.authorization.k8s.io/v1┌──[root@vms81.liruilongs.github.io]-[~/efk]└─$
复制代码
架构简述
关于搭建的架构。我们这里简单描述下，其实和我之前讲的 k8s 集群监控类似，下面为我们搭建好的EFK相关的pod列表，我们可以看到，elasticsearch集群作为日志存储平台，向上对接kibana(kibana是一个 Web 应用，用于调用接口查询elasticsearch集群的数据)，向下对接metricbeat，filebeat,这两个组件主要用于工作节点的日志采集,采集完日志给elasticsearch集群
┌──[root@vms81.liruilongs.github.io]-[~/ansible/k8s-helm-create/metricbeat]└─$kubectl get pods -o wideNAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES....elasticsearch-master-0 1/1 Running 2 (3h21m ago) 3h55m 10.244.70.40 vms83.liruilongs.github.io <none> <none>elasticsearch-master-1 1/1 Running 2 (3h21m ago) 3h55m 10.244.171.163 vms82.liruilongs.github.io <none> <none>filebeat-filebeat-64f5b 1/1 Running 7 (168m ago) 3h54m 10.244.70.36 vms83.liruilongs.github.io <none> <none>filebeat-filebeat-g4cw8 1/1 Running 8 (168m ago) 3h54m 10.244.171.185 vms82.liruilongs.github.io <none> <none>kibana-kibana-f88767f86-vqnms 1/1 Running 8 (167m ago) 3h57m 10.244.171.159 vms82.liruilongs.github.io <none> <none>........metricbeat-kube-state-metrics-75c5fc65d9-86fvh 1/1 Running 0 13m 10.244.70.43 vms83.liruilongs.github.io <none> <none>metricbeat-metricbeat-895qz 1/1 Running 0 13m 10.244.171.172 vms82.liruilongs.github.io <none> <none>metricbeat-metricbeat-metrics-7c5cd7d77f-22fgr 1/1 Running 0 13m 10.244.70.45 vms83.liruilongs.github.io <none> <none>metricbeat-metricbeat-n2gx2 1/1 Running 0 13m 10.244.70.46 vms83.liruilongs.github.io <none> <none>.........
复制代码
搭建环境
helm 版本
┌──[root@vms81.liruilongs.github.io]-[/var/lib/docker/containers/0606dd217f4f2f315f86fb378ffa65ed0c59ba7580869b9a88634dd6e171fdc0]└─$helm versionversion.BuildInfo{Version:"v3.2.1", GitCommit:"fe51cd1e31e6a202cba7dead9552a6d418ded79a", GitTreeState:"clean", GoVersion:"go1.13.10"}
复制代码
k8s 集群版本
┌──[root@vms81.liruilongs.github.io]-[/var/lib/docker/containers/0606dd217f4f2f315f86fb378ffa65ed0c59ba7580869b9a88634dd6e171fdc0]└─$kubectl get nodesNAME STATUS ROLES AGE VERSIONvms81.liruilongs.github.io Ready control-plane,master 54d v1.22.2vms82.liruilongs.github.io Ready <none> 54d v1.22.2vms83.liruilongs.github.io Ready <none> 54d v1.22.2
复制代码
EFK(helm）源设置
┌──[root@vms81.liruilongs.github.io]-[~/ansible/k8s-helm-create]└─$helm repo add elastic https://helm.elastic.co"elastic" has been added to your repositories
复制代码
┌──[root@vms81.liruilongs.github.io]-[~/ansible/k8s-helm-create]└─$helm repo listNAME URLazure http://mirror.azure.cn/kubernetes/charts/ali https://apphub.aliyuncs.comliruilong_repo http://192.168.26.83:8080/chartsstable https://charts.helm.sh/stableprometheus-community https://prometheus-community.github.io/helm-chartselastic https://helm.elastic.co
复制代码
安装的版本为： --version=7.9.1
EFK chart 包下载
┌──[root@vms81.liruilongs.github.io]-[~/ansible/k8s-helm-create]└─$helm pull elastic/elasticsearch --version=7.9.1┌──[root@vms81.liruilongs.github.io]-[~/ansible/k8s-helm-create]└─$helm pull elastic/filebeat --version=7.9.1┌──[root@vms81.liruilongs.github.io]-[~/ansible/k8s-helm-create]└─$helm pull elastic/metricbeat --version=7.9.1┌──[root@vms81.liruilongs.github.io]-[~/ansible/k8s-helm-create]└─$helm pull elastic/kibana --version=7.9.1
复制代码
查看 chart 下载列表，下载好之后直接通过 tar zxf 解压即可
┌──[root@vms81.liruilongs.github.io]-[~/ansible/k8s-helm-create]└─$ls *7.9.1*elasticsearch-7.9.1.tgz filebeat-7.9.1.tgz kibana-7.9.1.tgz metricbeat-7.9.1.tgz
复制代码
涉及到的相关镜像导入
需要注意的是，这里镜像下载特别费时间，最好提前下载一下，下面是我在所以工作节点导入的命令，要注意所以的节点都需要导入
┌──[root@vms82.liruilongs.github.io]-[/]└─$docker load -i elastic7.9.1.tar....Loaded image: docker.elastic.co/elasticsearch/elasticsearch:7.9.1┌──[root@vms82.liruilongs.github.io]-[/]└─$docker load -i filebeat7.9.1.tar....Loaded image: docker.elastic.co/beats/filebeat:7.9.1┌──[root@vms82.liruilongs.github.io]-[/]└─$docker load -i kibana7.9.1.tar....Loaded image: docker.elastic.co/kibana/kibana:7.9.1┌──[root@vms82.liruilongs.github.io]-[/]└─$docker load -i metricbeat7.9.1.tar.....Loaded image: docker.elastic.co/beats/metricbeat:7.9.1┌──[root@vms82.liruilongs.github.io]-[/]└─$
复制代码
EFK heml 安装
这里需要注意的是，我们还需要修改一些配置文件，根据集群的情况
elasticsearch 安装
es 集群参数的修改
┌──[root@vms81.liruilongs.github.io]-[~/ansible/k8s-helm-create]└─$cd elasticsearch/┌──[root@vms81.liruilongs.github.io]-[~/ansible/k8s-helm-create/elasticsearch]└─$lsChart.yaml examples Makefile README.md templates values.yaml
复制代码
修改集群副本数为 2，我们只有 2 个工作节点，所以只能部署两个
┌──[root@vms81.liruilongs.github.io]-[~/ansible/k8s-helm-create/elasticsearch]└─$cat values.yaml | grep replireplicas: 3 # # Add a template to adjust number of shards/replicas # curl -XPUT "$ES_URL/_template/$TEMPLATE_NAME" -H 'Content-Type: application/json' -d'{"index_patterns":['\""$INDEX_PATTERN"\"'],"settings":{"number_of_shards":'$SHARD_COUNT',"number_of_replicas":'$REPLICA_COUNT'}}'┌──[root@vms81.liruilongs.github.io]-[~/ansible/k8s-helm-create/elasticsearch]└─$sed 's#replicas: 3#replicas: 2#g' values.yaml | grep replicas:replicas: 2┌──[root@vms81.liruilongs.github.io]-[~/ansible/k8s-helm-create/elasticsearch]└─$sed -i 's#replicas: 3#replicas: 2#g' values.yaml
复制代码
修改集群最小 master 节点数为 1
┌──[root@vms81.liruilongs.github.io]-[~/ansible/k8s-helm-create/elasticsearch]└─$cat values.yaml | grep miniminimumMasterNodes: 2## ref: https://kubernetes.io/docs/tasks/administer-cluster/configure-multiple-schedulers/┌──[root@vms81.liruilongs.github.io]-[~/ansible/k8s-helm-create/elasticsearch]└─$sed -i 's#minimumMasterNodes: 2#minimumMasterNodes: 1#g' values.yaml
复制代码
修改数据持续化方式，这里修改不持久化
┌──[root@vms81.liruilongs.github.io]-[~/ansible/k8s-helm-create/elasticsearch]└─$cat values.yaml | grep -A 2 persistence:persistence: enabled: false labels:┌──[root@vms81.liruilongs.github.io]-[~/ansible/k8s-helm-create/elasticsearch]└─$
复制代码
安裝 elasticsearch
┌──[root@vms81.liruilongs.github.io]-[~/ansible/k8s-helm-create]└─$helm install elasticsearch elasticsearchNAME: elasticsearchLAST DEPLOYED: Sat Feb 5 03:15:45 2022NAMESPACE: kube-systemSTATUS: deployedREVISION: 1NOTES:1. Watch all cluster members come up. $ kubectl get pods --namespace=kube-system -l app=elasticsearch-master -w2. Test cluster health using Helm test. $ helm test elasticsearch --cleanup
复制代码
┌──[root@vms81.liruilongs.github.io]-[~/ansible/k8s-helm-create]└─$kubectl get pods --namespace=kube-system -l app=elasticsearch-master -wNAME READY STATUS RESTARTS AGEelasticsearch-master-0 0/1 Running 0 41selasticsearch-master-1 0/1 Running 0 41s
复制代码
等待一些时间，查看 pod 状态，elasticsearc 集群安装成功
┌──[root@vms81.liruilongs.github.io]-[~/ansible/k8s-helm-create]└─$kubectl get pods --namespace=kube-system -l app=elasticsearch-masterNAME READY STATUS RESTARTS AGEelasticsearch-master-0 1/1 Running 0 2m23selasticsearch-master-1 1/1 Running 0 2m23s┌──[root@vms81.liruilongs.github.io]-[~/ansible/k8s-helm-create]└─$
复制代码
metricbeat 安装
metricbeat 的安装就需要注意资源文件的版本问题,直接安装会包如下错误
报错问题解决
┌──[root@vms81.liruilongs.github.io]-[~/ansible/k8s-helm-create/metricbeat]└─$helm install metricbeat .Error: unable to build kubernetes objects from release manifest: [unable to recognize "": no matches for kind "ClusterRole" in version "rbac.authorization.k8s.io/v1beta1", unable to recognize "": no matches for kind "ClusterRoleBinding" in version "rbac.authorization.k8s.io/v1beta1"]
复制代码
解决办法，涉及到版本的直接修改
┌──[root@vms81.liruilongs.github.io]-[~/ansible/k8s-helm-create/metricbeat/charts/kube-state-metrics/templates]└─$sed -i 's#rbac.authorization.k8s.io/v1beta1#rbac.authorization.k8s.io/v1#g' *.yaml
复制代码
┌──[root@vms81.liruilongs.github.io]-[~/ansible/k8s-helm-create/metricbeat]└─$helm install metricbeat .NAME: metricbeatLAST DEPLOYED: Sat Feb 5 03:42:48 2022NAMESPACE: kube-systemSTATUS: deployedREVISION: 1TEST SUITE: NoneNOTES:1. Watch all containers come up. $ kubectl get pods --namespace=kube-system -l app=metricbeat-metricbeat -w
复制代码
┌──[root@vms81.liruilongs.github.io]-[~/ansible/k8s-helm-create/metricbeat]└─$kubectl get pods --namespace=kube-system -l app=metricbeat-metricbeatNAME READY STATUS RESTARTS AGEmetricbeat-metricbeat-cvqsm 1/1 Running 0 65smetricbeat-metricbeat-gfdqz 1/1 Running 0 65s
复制代码
filebeat 安装
安装 filebeat
┌──[root@vms81.liruilongs.github.io]-[~/ansible/k8s-helm-create/filebeat]└─$helm install filebeat .NAME: filebeatLAST DEPLOYED: Sat Feb 5 03:27:13 2022NAMESPACE: kube-systemSTATUS: deployedREVISION: 1TEST SUITE: NoneNOTES:1. Watch all containers come up. $ kubectl get pods --namespace=kube-system -l app=filebeat-filebeat -w
复制代码
过一会时间查看 pod 状态，安装成功
┌──[root@vms81.liruilongs.github.io]-[~/ansible/k8s-helm-create/filebeat]└─$kubectl get pods --namespace=kube-system -l app=filebeat-filebeat -wNAME READY STATUS RESTARTS AGEfilebeat-filebeat-df4s4 1/1 Running 0 20sfilebeat-filebeat-hw9xh 1/1 Running 0 21s
复制代码
kibana 安装
修改 SVC 类型
这里需要注意的是，我么需要找集群外部访问，所以 SVC 需要修改为NodePort
┌──[root@vms81.liruilongs.github.io]-[~/ansible/k8s-helm-create/kibana]└─$cat values.yaml | grep -A 2 "type: ClusterIP" type: ClusterIP loadBalancerIP: "" port: 5601┌──[root@vms81.liruilongs.github.io]-[~/ansible/k8s-helm-create/kibana]└─$sed -i 's#type: ClusterIP#type: NodePort#g' values.yaml┌──[root@vms81.liruilongs.github.io]-[~/ansible/k8s-helm-create/kibana]└─$cat values.yaml | grep -A 3 service:service: type: NodePort loadBalancerIP: "" port: 5601
复制代码
安装 kibana
┌──[root@vms81.liruilongs.github.io]-[~/ansible/k8s-helm-create/kibana]└─$helm install kibana .NAME: kibanaLAST DEPLOYED: Sat Feb 5 03:47:07 2022NAMESPACE: kube-systemSTATUS: deployedREVISION: 1TEST SUITE: None
复制代码
pending 解决
一直创建不成功，我们查看事件，发现是 CPU 和内存不够，没办法，这里只能重新调整虚机资源
┌──[root@vms81.liruilongs.github.io]-[~/ansible/k8s-helm-create/kibana]└─$kubectl get pods kibana-kibana-f88767f86-fsbhfNAME READY STATUS RESTARTS AGEkibana-kibana-f88767f86-fsbhf 0/1 Pending 0 6m14s┌──[root@vms81.liruilongs.github.io]-[~/ansible/k8s-helm-create/kibana]└─$kubectl describe pods kibana-kibana-f88767f86-fsbhf | grep -A 10 -i eventsEvents: Type Reason Age From Message ---- ------ ---- ---- ------- Warning FailedScheduling 4s (x7 over 6m33s) default-scheduler 0/3 nodes are available: 1 node(s) had taint {node-role.kubernetes.io/master: }, that the pod didn't tolerate, 2 Insufficient cpu, 2 Insufficient memory.┌──[root@vms81.liruilongs.github.io]-[~/ansible/k8s-helm-create/kibana]└─$
复制代码
调整完资源可能还会报这样的错，查了下，还是资源的问题，我的资源没办法调整了，后就休息了，睡起来发现可以了
Readiness probe failed: Error: Got HTTP code 000 but expected a 200
复制代码
┌──[root@vms81.liruilongs.github.io]-[~/ansible/k8s-helm-create]└─$kubectl describe pods kibana-kibana-f88767f86-qkrk7 | grep -i -A 10 eventEvents: Type Reason Age From Message ---- ------ ---- ---- ------- Normal Scheduled 8m21s default-scheduler Successfully assigned kube-system/kibana-kibana-f88767f86-qkrk7 to vms83.liruilongs.github.io Normal Pulled 7m42s kubelet Container image "docker.elastic.co/kibana/kibana:7.9.1" already present on machine Normal Created 7m42s kubelet Created container kibana Normal Started 7m41s kubelet Started container kibana Warning Unhealthy 4m17s (x19 over 7m30s) kubelet Readiness probe failed: Error: Got HTTP code 000 but expected a 200 Warning Unhealthy 2m19s (x6 over 7m5s) kubelet Readiness probe failed:┌──[root@vms81.liruilongs.github.io]-[~/ansible/k8s-helm-create]└─$
复制代码
下面是一些解决办法的参考，小伙伴有遇到可以看看
https://stackoverflow.com/questions/68831025/kibana-health-prob-fails-when-elasticsearch-host-is-added-to-fluentd-forwarder-c
https://stackoverflow.com/questions/48540929/kubernetes-readiness-probe-failed-error
通过 helm ls 可以看到我们的 EFK 已经安装完成了
┌──[root@vms81.liruilongs.github.io]-[~/ansible/k8s-helm-create]└─$helm lsNAME NAMESPACE REVISION UPDATED STATUS CHART APP VERSIONelasticsearch kube-system 1 2022-02-05 03:15:45.827750596 +0800 CST deployed elasticsearch-7.9.1 7.9.1filebeat kube-system 1 2022-02-05 03:27:13.473157636 +0800 CST deployed filebeat-7.9.1 7.9.1kibana kube-system 1 2022-02-05 03:47:07.618651858 +0800 CST deployed kibana-7.9.1 7.9.1metricbeat kube-system 1 2022-02-05 03:42:48.807772112 +0800 CST deployed metricbeat-7.9.1 7.9.1┌──[root@vms81.liruilongs.github.io]-[~/ansible/k8s-helm-create]└─$
复制代码
pod 列表查看所有的 pod 都 Running 了
┌──[root@vms81.liruilongs.github.io]-[~/ansible/k8s-helm-create/metricbeat]└─$kubectl get pods -o wideNAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATEScalico-kube-controllers-78d6f96c7b-85rv9 1/1 Running 327 (172m ago) 51d 10.244.88.83 vms81.liruilongs.github.io <none> <none>calico-node-6nfqv 1/1 Running 364 (172m ago) 54d 192.168.26.81 vms81.liruilongs.github.io <none> <none>calico-node-fv458 1/1 Running 70 (172m ago) 54d 192.168.26.83 vms83.liruilongs.github.io <none> <none>calico-node-h5lsq 1/1 Running 135 (3h25m ago) 54d 192.168.26.82 vms82.liruilongs.github.io <none> <none>coredns-7f6cbbb7b8-ncd2s 1/1 Running 33 (3h ago) 51d 10.244.88.84 vms81.liruilongs.github.io <none> <none>coredns-7f6cbbb7b8-pjnct 1/1 Running 30 (3h35m ago) 51d 10.244.88.82 vms81.liruilongs.github.io <none> <none>elasticsearch-master-0 1/1 Running 2 (3h21m ago) 3h55m 10.244.70.40 vms83.liruilongs.github.io <none> <none>elasticsearch-master-1 1/1 Running 2 (3h21m ago) 3h55m 10.244.171.163 vms82.liruilongs.github.io <none> <none>etcd-vms81.liruilongs.github.io 1/1 Running 134 (6h58m ago) 54d 192.168.26.81 vms81.liruilongs.github.io <none> <none>filebeat-filebeat-64f5b 1/1 Running 7 (168m ago) 3h54m 10.244.70.36 vms83.liruilongs.github.io <none> <none>filebeat-filebeat-g4cw8 1/1 Running 8 (168m ago) 3h54m 10.244.171.185 vms82.liruilongs.github.io <none> <none>kibana-kibana-f88767f86-vqnms 1/1 Running 8 (167m ago) 3h57m 10.244.171.159 vms82.liruilongs.github.io <none> <none>kube-apiserver-vms81.liruilongs.github.io 1/1 Running 32 (179m ago) 20d 192.168.26.81 vms81.liruilongs.github.io <none> <none>kube-controller-manager-vms81.liruilongs.github.io 1/1 Running 150 (170m ago) 52d 192.168.26.81 vms81.liruilongs.github.io <none> <none>kube-proxy-scs6x 1/1 Running 22 (3h21m ago) 54d 192.168.26.82 vms82.liruilongs.github.io <none> <none>kube-proxy-tbwz5 1/1 Running 30 (6h58m ago) 54d 192.168.26.81 vms81.liruilongs.github.io <none> <none>kube-proxy-xccmp 1/1 Running 11 (3h21m ago) 54d 192.168.26.83 vms83.liruilongs.github.io <none> <none>kube-scheduler-vms81.liruilongs.github.io 1/1 Running 302 (170m ago) 54d 192.168.26.81 vms81.liruilongs.github.io <none> <none>metricbeat-kube-state-metrics-75c5fc65d9-86fvh 1/1 Running 0 13m 10.244.70.43 vms83.liruilongs.github.io <none> <none>metricbeat-metricbeat-895qz 1/1 Running 0 13m 10.244.171.172 vms82.liruilongs.github.io <none> <none>metricbeat-metricbeat-metrics-7c5cd7d77f-22fgr 1/1 Running 0 13m 10.244.70.45 vms83.liruilongs.github.io <none> <none>metricbeat-metricbeat-n2gx2 1/1 Running 0 13m 10.244.70.46 vms83.liruilongs.github.io <none> <none>metrics-server-bcfb98c76-d58bg 1/1 Running 3 (3h13m ago) 6h28m 10.244.70.31 vms83.liruilongs.github.io <none> <none>
复制代码
查看节点核心指标
┌──[root@vms81.liruilongs.github.io]-[~/ansible/k8s-helm-create/metricbeat]└─$kubectl top nodesNAME CPU(cores) CPU% MEMORY(bytes) MEMORY%vms81.liruilongs.github.io 335m 16% 1443Mi 46%vms82.liruilongs.github.io 372m 12% 2727Mi 57%vms83.liruilongs.github.io 554m 13% 2513Mi 56%
复制代码
EFK 测试
查看 kibana 对应的端口，访问测试
┌──[root@vms81.liruilongs.github.io]-[~/ansible/k8s-helm-create/metricbeat]└─$kubectl get svcNAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGEelasticsearch-master ClusterIP 10.96.232.233 <none> 9200/TCP,9300/TCP 3h44melasticsearch-master-headless ClusterIP None <none> 9200/TCP,9300/TCP 3h44mkibana-kibana NodePort 10.100.170.124 <none> 5601:31335/TCP 3h46mkube-dns ClusterIP 10.96.0.10 <none> 53/UDP,53/TCP,9153/TCP 54dliruilong-kube-prometheus-kubelet ClusterIP None <none> 10250/TCP,10255/TCP,4194/TCP 20dmetricbeat-kube-state-metrics ClusterIP 10.108.35.197 <none> 8080/TCP 115smetrics-server ClusterIP 10.111.104.173 <none> 443/TCP 54dnginxdep NodePort 10.106.217.50 <none> 8888:31964/TCP 2d20h
复制代码
评论