helm部署Filebeat + ELK

系统架构图：

1) 多个Filebeat在各个Node进行日志采集，然后上传至Logstash

2) 多个Logstash节点并行（负载均衡，不作为集群），对日志记录进行过滤处理，然后上传至Elasticsearch集群

3) 多个Elasticsearch构成集群服务，提供日志的索引和存储能力

4) Kibana负责对Elasticsearch中的日志数据进行检索、分析

1. Elasticsearch部署

官方chart地址：https://github.com/elastic/helm-charts/tree/master/elasticsearch

创建logs命名空间

1	`kubectl create ns logs`

添加elastic helm charts 仓库

1	`helm repo add elastic https://helm.elastic.co`

安装

1	`helm` `install` `--name elasticsearch elastic/elasticsearch` `--namespace logs`

参数说明

image: "docker.elastic.co/elasticsearch/elasticsearch"
 
imageTag: "7.2.0"
 
imagePullPolicy: "IfNotPresent"
 
podAnnotations: {}
 
esJavaOpts: "-Xmx1g -Xms1g"
 
resources:
 
  requests:
 
    cpu: "100m"
 
    memory: "2Gi"
 
  limits:
 
    cpu: "1000m"
 
    memory: "2Gi"
 
volumeClaimTemplate:
 
  accessModes: [ "ReadWriteOnce" ]
 
  storageClassName: "nfs-client"
 
  resources:
 
    requests:
 
      storage: 50Gi

2. Filebeat部署

官方chart地址：https://github.com/elastic/helm-charts/tree/master/filebeat

Add the elastic helm charts repo

1	`helm repo add elastic https://helm.elastic.co`

Install it

1	`helm` `install` `--name filebeat elastic/filebeat` `--namespace logs`

参数说明：

image: "docker.elastic.co/beats/filebeat"
 
imageTag: "7.2.0"
 
imagePullPolicy: "IfNotPresent"
 
resources:
 
  requests:
 
    cpu: "100m"
 
    memory: "100Mi"
 
  limits:
 
    cpu: "1000m"
 
    memory: "200Mi"

那么问题来了，filebeat默认收集宿主机上docker的日志路径：/var/lib/docker/containers。如果我们修改了docker的安装路径要怎么收集呢，很简单修改chart里的DaemonSet文件里边的hostPath参数：

- name: varlibdockercontainers
  hostPath:
    path: /var/lib/docker/containers   #改为docker安装路径

对java程序的报错异常log实现多行合并，用multiline定义正则来匹配。

filebeatConfig:
  filebeat.yml: |
    filebeat.inputs:
    - type: docker
      containers.ids:
      - '*'
      multiline.pattern: '^[0-9]'
      multiline.negate: true
      multiline.match: after
      processors:
      - add_kubernetes_metadata:
          in_cluster: true
 
    output.elasticsearch:
      hosts: '${ELASTICSEARCH_HOSTS:elasticsearch-master:9200}'

3. Kibana部署

官方chart地址：https://github.com/elastic/helm-charts/tree/master/kibana

Add the elastic helm charts repo

1	`helm repo add elastic https://helm.elastic.co`

Install it

1	`helm` `install` `--name kibana elastic/kibana` `--namespace logs`

参数说明：

elasticsearchHosts: "http://elasticsearch-master:9200"
 
replicas: 1
 
image: "docker.elastic.co/kibana/kibana"
 
imageTag: "7.2.0"
 
imagePullPolicy: "IfNotPresent"
 
resources:
 
  requests:
 
    cpu: "100m"
 
    memory: "500m"
 
  limits:
 
    cpu: "1000m"
 
    memory: "1Gi"

4. Logstash部署

官方chart地址：https://github.com/helm/charts/tree/master/stable/logstash

安装

1	`$ helm` `install` `--name logstash stable/logstash` `--namespace logs`

参数说明：

image:
 
  repository: docker.elastic.co/logstash/logstash-oss
 
  tag: 7.2.0
 
  pullPolicy: IfNotPresent
 
persistence:
 
  enabled: true
 
  storageClass: "nfs-client"
 
  accessMode: ReadWriteOnce
 
  size: 2Gi

匹配label：json的pod日志，没有的话正常收集。

filebeatConfig:
  filebeat.yml: |
    filebeat.autodiscover:
      providers:
        - type: kubernetes
          templates:
            - condition:
                equals:
                  kubernetes.labels.logFormat: "json"
              config:
                - type: docker
                  containers.ids:
                    - "${data.kubernetes.container.id}"
                  json.keys_under_root: true
                  json.overwrite_keys: true
                  json.add_error_key: true
            - config:
                - type: docker
                  containers.ids:
                    - "${data.kubernetes.container.id}"
                  processors:
                    - add_kubernetes_metadata:
                        in_cluster: true
    output.elasticsearch:
      hosts: '${ELASTICSEARCH_HOSTS:elasticsearch-master:9200}'