helm部署skywalking链路追踪 java

添加helm仓库 skywalking取别名 sw 名称可以任意写helm repo add sw https://apache.jfrog.io/artifactory/skywalking-helm helm repo list这里 sw 要与上面的 sw 名称 一样从 Helm 仓库下载 SkyWalking 的 Chart 包–untar 并自动解压到当前目录helm pull sw/skywalking--untar进目录cd skywalking/测试环境用的配置vim sky.yaml# OAP 配置oap:image:repository:apache/skywalking-oap-servertag:9.7.0pullPolicy:IfNotPresentstorageType:elasticsearchreplicas:1# 单副本数javaOpts:-Xmx1g -Xms1ginitEs:trueservice:type:ClusterIP# UI 配置ui:image:repository:apache/skywalking-uitag:9.7.0pullPolicy:IfNotPresentreplicas:1service:type:NodePortexternalPort:80internalPort:8080# Elasticsearch 配置使用内置 ESelasticsearch:enabled:trueimageTag:7.17.3# 使用 ES 7.xSkyWalking 9.x 兼容性好replicas:1# 单节点测试persistence:enabled:false# 环境不持久化# 资源配置resources:requests:cpu:500mmemory:1Gilimits:cpu:1000mmemory:2Gi# 单节点配置minimumMasterNodes:1roles:master:truedata:trueingest:true# 禁用其他功能satellite:enabled:falseelasticsearch6:enabled:falsefullnameOverride:nameOverride:生产环境用的配置vim sky.yaml# OAP 配置oap:image:repository:apache/skywalking-oap-servertag:9.7.0pullPolicy:IfNotPresentstorageType:elasticsearchreplicas:2# OAP 可以扩展到 2 个实例javaOpts:-Xmx1g -Xms1ginitEs:trueservice:type:ClusterIP# 资源限制resources:requests:memory:2Gi# 给 JVM 堆外内存留空间cpu:500mlimits:memory:3Gicpu:1000m# 存活探针防止僵尸进程livenessProbe:httpGet:path:/healthport:12800initialDelaySeconds:60periodSeconds:10# 就绪探针控制流量readinessProbe:httpGet:path:/healthport:12800initialDelaySeconds:30periodSeconds:5# UI 配置ui:image:repository:apache/skywalking-uitag:9.7.0pullPolicy:IfNotPresentreplicas:1service:type:NodePortexternalPort:80internalPort:8080# 资源限制resources:requests:memory:500Micpu:100mlimits:memory:1024Micpu:500m# 探针配置livenessProbe:httpGet:path:/healthport:8080initialDelaySeconds:30periodSeconds:10# Elasticsearch 3节点集群配置elasticsearch:enabled:trueimageTag:7.17.3replicas:3# 持久化存储重要persistence:enabled:truevolumeClaimTemplate:accessModes:-ReadWriteOncestorageClassName:csi-diskresources:requests:storage:30Gi# 资源配置resources:requests:cpu:500mmemory:2Gilimits:cpu:1000mmemory:3Gi# 集群配置minimumMasterNodes:2roles:master:truedata:trueingest:true# 调度策略antiAffinity:soft# 尽量分散到不同节点nodeAffinity:{}# 可指定节点标签tolerations:[]# 可添加容忍度# 性能优化esJavaOpts:-Xmx1g -Xms1gsysctlVmMaxMapCount:262144sysctlInitContainer:enabled:true# 服务配置service:type:ClusterIPhttpPort:9200transportPort:9300# 禁用其他功能satellite:enabled:falseelasticsearch6:enabled:falsefullnameOverride:nameOverride:安装 skywalking下面示例是用的 生产的 sky.yaml 配置直接运行会报错 先替换掉所有的 v1beta1grep-r v1beta1 find.-typef-exec sed-is/v1beta1/v1/g{}\;helm install skywalking.--namespace skywalking--create-namespace-f sky.yaml就可以通过 节点ip端口访问页面客户端deploy yaml 的配置要实现收集数据 需要修改 yaml 主要是改下面3段如果配置以后启动不了 看下是否是 limit的cpu限制导致的 可以适当调大limit下面是3段参数volumes:-name:skywalking-agentemptyDir:{}initContainers:-name:sw-agent-downloadimage:apache/skywalking-java-agent:9.6.0-java17command:[cp,-r,/skywalking/agent,/target/]volumeMounts:-name:skywalking-agentmountPath:/target-name:JAVA_TOOL_OPTIONSvalue:-javaagent:/skywalking/agent/skywalking-agent.jar -Dskywalking.agent.service_namedemo-app -Dskywalking.collector.backend_serviceskywalking-oap.skywalking.svc.cluster.local:11800volumeMounts:-name:skywalking-agentmountPath:/skywalking