自拍偷在线精品自拍偷,亚洲欧美中文日韩v在线观看不卡

ELK Stack生產(chǎn)實(shí)踐——pod日志采集(EFK方案)

開(kāi)發(fā) 架構(gòu)
默認(rèn)情況下fluent-bit通過(guò)socket獲取journald日志,但如果使用pod運(yùn)行fluent-bit時(shí)無(wú)法獲取journald日志,需要配置journald參數(shù),改為存儲(chǔ)至宿主機(jī)本地/var/log/journal目錄下,然后pod掛載宿主機(jī)/var/log目錄進(jìn)而讀取kubelet和container日志。

EFK日志采集方案

方案介紹

  • 實(shí)現(xiàn)目標(biāo):k8s集群所有節(jié)點(diǎn)container和kubelet服務(wù)日志采集、k8s集群所有pod日志采集并對(duì)自定義應(yīng)用日志實(shí)現(xiàn)過(guò)濾和清洗操作。
  • fluent bit采集:相較于fluentd而言,fluent bit更加輕量,且內(nèi)置了pod日志解析插件和service日志采集功能無(wú)需編寫復(fù)雜的日志處理配置。因此使用daemonset方式在每個(gè)節(jié)點(diǎn)部署fluent bit容器,可以最大程度減少資源占用。
  • fluentd處理:雖然fluent bit足夠輕量,但日志解析處理能力弱于fluentd。fluentd支持更多的過(guò)濾處理插件,對(duì)于常用的數(shù)據(jù)處理操作,fluentd可以直接通過(guò)ruby語(yǔ)法處理,而fluent bit需要開(kāi)發(fā)lua腳本才能實(shí)現(xiàn)。
  • fluentd集群方案:使用Fluentd充當(dāng)日志聚合層,接收f(shuō)luent-bit日志后統(tǒng)一進(jìn)行處理操作,最后批量寫入elasticsearch集群。這樣做的好處是當(dāng)k8s集群規(guī)模過(guò)大時(shí),避免了過(guò)多的fluent-bit連接ES寫入數(shù)據(jù),導(dǎo)致ES連接資源消耗過(guò)高、網(wǎng)絡(luò)擁堵、連接資源競(jìng)爭(zhēng)等問(wèn)題。
  • fluent bit+fluentd方案:將日志采集與日志處理拆后分別交由不同的組件負(fù)責(zé),最大程度發(fā)揮各個(gè)組件的優(yōu)勢(shì),使得配置文件更加清晰易讀,便于后期維護(hù)管理。

架構(gòu)圖

圖片圖片

fluent bit部署與配置

參考文檔

https://docs.fluentbit.io/manual/installation/kubernetes

開(kāi)啟journal日志持久化

默認(rèn)情況下fluent-bit通過(guò)socket獲取journald日志,但如果使用pod運(yùn)行fluent-bit時(shí)無(wú)法獲取journald日志,需要配置journald參數(shù),改為存儲(chǔ)至宿主機(jī)本地/var/log/journal目錄下,然后pod掛載宿主機(jī)/var/log目錄進(jìn)而讀取kubelet和container日志。

[root@work1 ~]# vim /etc/systemd/journald.conf
Storage=persistent
[root@work1 ~]# systemctl restart systemd-journald

部署fluent-bit

拉取helm包

[root@master1 ~]# helm repo add fluent https://fluent.github.io/helm-charts
[root@master1 ~]# helm pull fluent/fluent-bit --untar
[root@master1 ~]# cd fluent-bit/
[root@master1 fluent-bit]# ls
Chart.yaml  ci  dashboards  README.md  templates  values.yaml

修改配置

[root@tiaoban fluent-bit]# vim values.yaml
image:
  repository: harbor.local.com/elk/fluentbit # 修改為本地倉(cāng)庫(kù)地址
  tag: v2.1.9 # 指定鏡像版本

config:
 # 默認(rèn)未配置文件緩沖,新增文件緩沖配置
  service: |
    [SERVICE]
        Daemon Off
        Flush {{ .Values.flush }}
        Log_Level {{ .Values.logLevel }}
        Parsers_File /fluent-bit/etc/parsers.conf
        Parsers_File /fluent-bit/etc/conf/custom_parsers.conf
        HTTP_Server On
        HTTP_Listen 0.0.0.0
        HTTP_Port {{ .Values.metricsPort }}
        Health_Check On
        storage.path              /var/log/flb-storage/ 
        storage.sync              normal               
        storage.checksum          off                   
        storage.backlog.mem_limit 5M 
 
  # 默認(rèn)配置文件只采集pod和kubelet服務(wù)日志,新增containerd日志,并排除Fluentd和Fluent-bit日志采集
  inputs: |
    [INPUT]
        Name systemd
        Tag containerd.service
        Systemd_Filter _SYSTEMD_UNIT=containerd.service
        Read_From_Tail On

    [INPUT]
        Name systemd
        Tag kubelet.service
        Systemd_Filter _SYSTEMD_UNIT=kubelet.service
        Read_From_Tail On

    [INPUT]
        Name tail
        Path /var/log/containers/*.log
        Exclude_Path /var/log/containers/fluent-bit*, /var/log/containers/fluentd-*
        multiline.parser docker, cri
        Tag kube.*
        Mem_Buf_Limit 5MB
        Skip_Long_Lines On
  
  # 原始字段中的time值不是標(biāo)準(zhǔn)時(shí)間格式,直接寫入es會(huì)報(bào)錯(cuò),可移除time字段,使用@timestamp標(biāo)識(shí)時(shí)間
 filters: |
    [FILTER]
        name        record_modifier
        match       kube.*
        remove_key  time
 
  # 將所有采集到的數(shù)據(jù)直接轉(zhuǎn)發(fā)至后端fluentd服務(wù)。
  outputs: |
    [OUTPUT]
        Name forward
        Match *
        Host fluentd.logging.svc
        Port 8888
        storage.total_limit_size 5M

hotReload:
  enabled: true # 按需開(kāi)啟配置熱更新
  image:
    repository: harbor.local.com/elk/fluentbit # 修改為本地倉(cāng)庫(kù)地址

部署fluent-bit

[root@tiaoban fluent-bit]# kubectl create ns logging
namespace/logging created
[root@tiaoban fluent-bit]# helm install fluent-bit -n logging . -f values.yaml
NAME: fluent-bit
LAST DEPLOYED: Sat Sep 16 14:46:03 2023
NAMESPACE: logging
STATUS: deployed
REVISION: 1
NOTES:
Get Fluent Bit build information by running these commands:

export POD_NAME=$(kubectl get pods --namespace logging -l "app.kubernetes.io/name=fluent-bit,app.kubernetes.io/instance=fluent-bit" -o jsnotallow="{.items[0].metadata.name}")
kubectl --namespace logging port-forward $POD_NAME 2020:2020
curl http://127.0.0.1:2020

驗(yàn)證

[root@tiaoban fluent]# cat Dockerfile 
FROM fluent/fluentd:v1.16.2-debian-1.1
USER root
ADD GeoLite2-City.mmdb /fluentd/GeoLite2-City.mmdb
RUN buildDeps="build-essential libgeoip-dev autoconf automake libtool libffi-dev" \
    && sed -i "s@http://deb.debian.org@http://mirrors.aliyun.com@g" /etc/apt/sources.list \
    && apt-get update \
    && apt-get install -y --no-install-recommends $buildDeps \
    && gem sources --add https://gems.ruby-china.com/ --remove https://rubygems.org/ \
    && fluent-gem install fluent-plugin-elasticsearch fluent-plugin-geoip fluent-plugin-rewrite-tag-filter \
    && apt-get autoclean \
    && gem sources --clear-all \
    && rm -rf /var/lib/apt/lists/* \
    && rm -rf /tmp/* /var/tmp/* /usr/lib/ruby/gems/*/cache/*.gem
[root@tiaoban fluent]# docker build -t harbor.local.com/elk/fluentd:fluentd:v1.16.2 .
[root@tiaoban fluent]# docker push harbor.local.com/elk/fluentd:fluentd:v1.16.2

fluentd部署與配置

構(gòu)建fluentd鏡像

由于fluentd鏡像未安裝elasticsearch、geoip、rewrite-tag-filter插件,如果在pod的init階段在線安裝插件經(jīng)常會(huì)導(dǎo)致拉取超時(shí)無(wú)法正常啟動(dòng)。因此需要提前構(gòu)建包含相關(guān)插件的fluentd鏡像,并上傳至harbor倉(cāng)庫(kù)中。

[root@tiaoban fluentd]# cat fluentd-conf.yaml 
apiVersion: v1
kind: ConfigMap
metadata:
  name: fluentd-conf
  namespace: logging
data:
  fluent.conf : |-
    <source>
      @type forward
      port 8888
      bind 0.0.0.0
    </source>

    <match *.service>
      @type elasticsearch
      host elasticsearch-es-http.elk.svc
      port 9200
      default_elasticsearch_version 8
      user elastic
      scheme https
      password 78HOWor95Iiot076O59xq2Am
      ssl_verify false
      data_stream_name logs-${tag}-fluentd
      include_timestamp true
      <buffer>
        @type file
        flush_interval 5s
        path /fluentd/buf/service-logs.*
      </buffer>
    </match>

    <match kube.**>
      @type elasticsearch
      host elasticsearch-es-http.elk.svc
      port 9200
      default_elasticsearch_version 8
      user elastic
      scheme https
      password 78HOWor95Iiot076O59xq2Am
      ssl_verify false
      data_stream_name logs-pod-fluentd
      include_timestamp true
      <buffer>
        @type file
        flush_interval 5s
        path /fluentd/buf/pod-logs.*
      </buffer>
    </match>

fluentd部署配置

fluentd使用deployment控制器部署多副本集群,用于接收f(shuō)luent bit的數(shù)據(jù)然后寫入es的數(shù)據(jù)流中。

  • fluentd-conf.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
  name: fluentd
  namespace: logging
spec:
  replicas: 3
  selector:
    matchLabels:
      app: fluentd
  template:
    metadata:
      labels:
        app: fluentd
    spec:
      containers:
      - name: fluentd
        image: harbor.local.com/elk/fluentd:v1.16.2
        resources:
          limits:
            memory: "128Mi"
            cpu: "500m"
        ports:
        - containerPort: 8888
        volumeMounts:
          - name: fluentd-conf
            mountPath: /fluentd/etc/fluent.conf
            subPath: fluent.conf
      volumes:
        - name: fluentd-conf
          configMap:
            name: fluentd-conf
  • fluentd-svc.yaml
apiVersion: v1
kind: Service
metadata:
  name: fluentd
  namespace: logging
spec:
  selector:
    app: fluentd
  ports:
  - port: 8888
    targetPort: 8888
  • fluentd.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
  name: fluentd
  namespace: logging
spec:
  replicas: 3
  selector:
    matchLabels:
      app: fluentd
  template:
    metadata:
      labels:
        app: fluentd
    spec:
      containers:
      - name: fluentd
        image: harbor.local.com/elk/fluentd:v1.16.2
        resources:
          limits:
            memory: "128Mi"
            cpu: "500m"
        ports:
        - containerPort: 8888
        volumeMounts:
          - name: fluentd-conf
            mountPath: /fluentd/etc/fluent.conf
            subPath: fluent.conf
      volumes:
        - name: fluentd-conf
          configMap:
            name: fluentd-conf

部署fluentd服務(wù)

[root@tiaoban fluentd]# ls
Dockerfile  fluentd-conf.yaml  fluentd-svc.yaml  fluentd.yaml  GeoLite2-City.mmdb
[root@tiaoban fluentd]# kubectl apply -f .
configmap/fluentd-conf created
service/fluentd created
deployment.apps/fluentd created

訪問(wèn)驗(yàn)證

此時(shí)查看kibana索引管理,已成功創(chuàng)建了pod、containerd、kubelet日志數(shù)據(jù)流。

圖片圖片

service日志處理

雖然我們已經(jīng)將service日志存儲(chǔ)在了elasticsearch中,但觀察發(fā)現(xiàn)默認(rèn)情況下日志內(nèi)容都在MESSAGE字段中,通常情況下,我們只會(huì)關(guān)注日志等級(jí)為warning和error的日志,因此我們需要對(duì)該字段內(nèi)容做解析處理。

日志內(nèi)容分析

觀察container日志可知,MESSAGE中包含了日志時(shí)間、日志等級(jí)、日志內(nèi)容三部分內(nèi)容

圖片圖片

觀察kubelet日志,MESSAGE中包含了日志時(shí)間、處理函數(shù)、日志內(nèi)容三部分內(nèi)容

圖片圖片

新增service日志處理規(guī)則

我們可以通過(guò)正則匹配插件實(shí)現(xiàn)MESSAGE字段值解析操作,需要注意的是正則匹配時(shí)一定是所有字段都包含,否則會(huì)因?yàn)槠ヅ洳坏蕉鵁o(wú)法過(guò)濾處理操作。

  • service-conf.yaml
apiVersion: v1
kind: ConfigMap
metadata:
  name: service-conf
  namespace: logging
data:
  service.conf : |
    <filter containerd.service>
      @type parser
      key_name MESSAGE
      reserve_data true
      <parse>
        @type regexp
        expression /time="(?<time>[^"]+)" level=(?<level>[^ ]+) msg="(?<msg>[^"]+)"/
        time_key time
        time_format %Y-%m-%dT%H:%M:%S.%N%z
      </parse>
    </filter>

    <filter kubelet.service>
      @type parser
      key_name MESSAGE
      reserve_data true
      <parse>
        @type regexp
        expression /[EI](?<time>\d{4} \d{2}:\d{2}:\d{2}.\d{6}) +\d* (?<class>\w*).go:\d*] (?<message>.*)/
        time_key time
        time_format %m%d %H:%M:%S.%N
        timezone +08:00
      </parse>
    </filter>

    <match *.service>
      @type elasticsearch_data_stream
      host elasticsearch-es-http.elk.svc
      port 9200
      default_elasticsearch_version 8
      user elastic
      scheme https
      password 78HOWor95Iiot076O59xq2Am
      ssl_verify false
      data_stream_name logs-${tag}-fluentd
      include_timestamp true
      <buffer>
        @type file
        path /fluentd/buf/service-logs.*
      </buffer>
    </match>

訪問(wèn)驗(yàn)證

查看containerd服務(wù)日志,已成功解析了MESSAGE字段內(nèi)容,并新增了@timestamp、leve、msg字段

圖片圖片

查看kubelet服務(wù)日志,同樣成功解析了MESSAGE字段內(nèi)容,并新增了@timestamp、class、message字段

圖片圖片

自定義日志解析

需求分析

默認(rèn)情況下,fluent bit會(huì)采集所有pod日志信息,并自動(dòng)添加namespace、pod、container等信息,所有日志內(nèi)容存儲(chǔ)在log字段中。 以log-demo應(yīng)為日志為例,將所有日志內(nèi)容存儲(chǔ)到log字段下無(wú)法很好的解析日志內(nèi)容,因此需要編寫fluentd解析規(guī)則,實(shí)現(xiàn)日志內(nèi)容解析。

圖片圖片

新增自定義日志處理規(guī)則

使用rewrite_tag_filter插件,匹配到container_name字段值為log-demo的事件時(shí),新增myapp.log標(biāo)簽,接下來(lái)對(duì)myapp.log標(biāo)簽的事件進(jìn)一步做解析處理操作。

  • myapp-conf.yaml
apiVersion: v1
kind: ConfigMap
metadata:
  name: myapp-conf
  namespace: logging
data:
  myapp.conf : |
    <match kube.**>
      @type rewrite_tag_filter
      <rule>
        key $.kubernetes.container_name
        pattern log-demo
        tag myapp.log
      </rule>
    </match>

    <filter myapp.log>
      @type parser
      key_name log
      <parse>
        @type regexp
        expression /^(?<logtime>[^|]+) \| (?<level>[A-Z]*) *\| __main__:(?<class>\D*:\d*) - (?<content>.*)$/
        time_key logtime
        time_format %Y-%m-%d %H:%M:%S.%L
      </parse>
    </filter>

    <filter myapp.log>
      @type record_transformer
      enable_ruby
      <record>
        content ${record["content"].gsub("'", '"')}
      </record>
    </filter>

    <filter myapp.log>
      @type parser
      key_name content
      <parse>
        @type json
      </parse>
    </filter>

    <filter myapp.log>
      @type geoip
      geoip_lookup_keys remote_address
      geoip2_database /fluentd/GeoLite2-City.mmdb
      backend_library geoip2_c
      <record>
        geoip_city            ${city.names.en["remote_address"]}
        geoip_latitude        ${location.latitude["remote_address"]}
        geoip_longitude       ${location.longitude["remote_address"]}
        geoip_country         ${country.iso_code["remote_address"]}
        geoip_country_name    ${country.names.en["remote_address"]}
        geoip_postal_code     ${postal.code["remote_address"]}
        geoip_region_name     ${subdivisions.0.names.en["remote_address"]}
      </record>
    </filter>

    <match myapp.log>
      @type elasticsearch_data_stream
      host elasticsearch-es-http.elk.svc
      port 9200
      default_elasticsearch_version 8
      user elastic
      scheme https
      password 78HOWor95Iiot076O59xq2Am
      ssl_verify false
      data_stream_name logs-myapp-default
      include_timestamp true
      <buffer>
        @type file
        flush_interval 5s
        path /fluentd/buf/myapp-logs.*
      </buffer>
    </match>

訪問(wèn)驗(yàn)證

查看kibana索引信息,已成功創(chuàng)建名為logs-myapp-default的數(shù)據(jù)流。

圖片圖片

查看logs-myapp-default字段信息,已成功解析了日志內(nèi)容。

圖片圖片

完整資源清單

本實(shí)驗(yàn)案例所有yaml文件已上傳至git倉(cāng)庫(kù)。訪問(wèn)地址如下:

Github

https://github.com/cuiliang0302/blog-demo

Gitee

https://gitee.com/cuiliang0302/blog_demo

參考文檔

fluent bit轉(zhuǎn)發(fā)事件至fluentd:

https://docs.fluentbit.io/manual/pipeline/outputs/forward

構(gòu)建自定義fluentd鏡像:

https://github.com/fluent/fluentd-docker-image#31-for-current-images

fluentd重寫事件tag:

https://docs.fluentd.org/output/rewrite_tag_filter

博客鏈接:https://www.cuiliangblog.cn/

責(zé)任編輯:武曉燕 來(lái)源: 崔亮的博客
相關(guān)推薦

2023-10-20 15:08:28

pod日志采集

2023-09-05 07:40:37

PythonSDKAPI

2017-01-10 16:04:02

容器MySQL實(shí)踐

2022-06-03 09:21:47

Svelte前端攜程

2020-12-31 10:48:17

ELKEFKLinux

2021-03-12 07:47:44

KubernetesRedis-clustRedis

2023-07-12 16:07:50

鏈路數(shù)據(jù)湖技術(shù)

2021-07-20 08:32:16

Kubernetes日志平臺(tái)

2021-07-19 09:18:07

KubernetesELK Stackk8s

2023-12-08 07:59:04

2023-10-09 14:36:28

工具PLGEFK

2023-10-16 16:00:27

Redis限流

2022-07-08 09:26:45

Flink快手計(jì)算

2022-03-08 08:21:21

Spring日志分析系統(tǒng)日志數(shù)據(jù)

2022-09-19 08:35:28

Kafka節(jié)點(diǎn)故障

2022-04-29 09:04:35

日志平臺(tái)開(kāi)發(fā)

2022-11-23 21:51:42

大數(shù)據(jù)

2024-12-09 08:27:02

2018-10-17 10:49:49

Kubernetes存儲(chǔ)處理
點(diǎn)贊
收藏

51CTO技術(shù)棧公眾號(hào)