ELK分析k8s应用日志
ELK分析k8s应用日志日志收集过程架构安装elasticsearch-6.8.6安装kafka安装filebeat日志收集过程收集,能够采集多种来源的日志数据(流式日志收集齐)传输,能够稳定的吧日志传输到中央系统;ElasticSearch可以通过9200http方式传输,也可以通过架构日志收集服务使用filebeat,因为传统的ruby收集的日志的方式十分消耗资源;其与业务服务采用边车模式运行
ELK分析k8s应用日志
日志收集过程
收集,能够采集多种来源的日志数据(流式日志收集齐)
传输,能够稳定的吧日志传输到中央系统;ElasticSearch可以通过9200http方式传输,也可以通过
架构
安装elasticsearch-6.8.6
服务部署在21上
src]# wget https://artifacts.elastic.co/downloads/elasticsearch/elasticsearch-6.8.6.tar.gz
src]# tar -xf elasticsearch-6.8.6.tar.gz -C /opt/release/
src]# ln -s /opt/release/elasticsearch-6.8.6 /opt/apps/elasticsearch
~]# grep -Ev "^$|^#" /opt/apps/elasticsearch/config/elasticsearch.yml # 调整以下配置
cluster.name: elasticsearch.host.com
path.data: /data/elasticsearch/data
path.logs: /data/elasticsearch/logs
bootstrap.memory_lock: true
network.host: 172.16.0.21
http.port: 9200
~]# vim /opt/apps/elasticsearch/config/jvm.options # 默认1g
......
# 生产中一般不超过32G
-Xms16g
-Xmx16g
~]# useradd -M es
~]# mkdir -p /data/elasticsearch/{logs,data}
~]# chown -R es.es /data/elasticsearch /opt/release/elasticsearch-6.8.6
## 修改es用户的内核配置
~]# vim /etc/security/limits.conf
......
es hard nofile 65536
es soft fsize unlimited
es hard memlock unlimited
es soft memlock unlimited
~]# echo "vm.max_map_count=262144" >> /etc/sysctl.conf ; sysctl -p
# 管理es命令
~]# su es -c "/opt/apps/elasticsearch/bin/elasticsearch -d" # 启动
~]# su es -c "/opt/apps/elasticsearch/bin/elasticsearch -d -p /data/elasticsearch/logs/pid" # 指定pid记录的文件启动es
~]# netstat -lntp | grep 9.00
tcp6 0 0 10.4.7.12:9200 :::* LISTEN 69352/java
tcp6 0 0 10.4.7.12:9300 :::* LISTEN 69352/java
~]# su es -c "ps aux|grep -v grep|grep java|grep elasticsearch|awk '{print \$2}'|xargs kill" # kill Pid
~]# pkill -F /data/elasticsearch/logs/pid
# 添加k8s日志索引模板
~]# curl -H "Content-Type:application/json" -XPUT http://172.16.0.21:9200/_template/k8s -d '{
"template" : "k8s*",
"index_patterns": ["k8s*"],
"settings": {
"number_of_shards": 5,
"number_of_replicas": 0
}
}'
安装kafka
最好不要超过2.2版本,超过之后kafka_manager就不再支持
src]# wget https://archive.apache.org/dist/kafka/2.2.0/kafka_2.12-2.2.0.tgz
src]# tar -xf kafka_2.12-2.2.0.tgz -C /opt/release/
src]# ln -s /opt/release/kafka_2.12-2.2.0 /opt/apps/kafka
~]# vim /opt/apps/kafka/config/server.properties
~]# vim /opt/apps/kafka/config/server.properties
......
log.dirs=/data/kafka/logs
# 超过10000条日志强制刷盘,超过1000ms刷盘
log.flush.interval.messages=10000
log.flush.interval.ms=1000
# 填写需要连接的 zookeeper 集群地址,当前连接本地的 zk 集群。
zookeeper.connect=localhost:2181
# 新增以下两项
delete.topic.enable=true
listeners=PLAINTEXT://172.16.0.21:9092
# 启动kafka之前需要先启动zookeeper
/opt/apps/kafka/bin/zookeeper-server-start.sh -daemon /opt/apps/kafka/config/zookeeper.properties
~]# mkdir -p /data/kafka/logs
~]# /opt/apps/kafka/bin/kafka-server-start.sh -daemon /opt/apps/kafka/config/server.properties
~]# netstat -lntp|grep 121952
tcp6 0 0 10.4.7.11:9092 :::* LISTEN 121952/java
tcp6 0 0 :::41211 :::* LISTEN 121952/java
# 查询kafka的topic是否创建成功
/opt/apps/kafka/bin/kafka-topics.sh --list --zookeeper localhost:2181
kafka-manager 安装
配置dockerfile
kafka-manager 改名为 CMAK,压缩包名称和内部目录名发生了变化,后续安装
# 这里存在几个问题:
# 1. kafka-manager 改名为 CMAK,压缩包名称和内部目录名发生了变化
# 2. sbt 编译需要下载很多依赖,因为不可描述的原因,速度非常慢,个人非VPN网络大概率失败
# 3. 因本人不具备VPN条件,编译失败。又因为第一条,这个dockerfile大概率需要修改
# 4. 生产环境中一定要自己重新做一份!
FROM hseeberger/scala-sbt
ENV ZK_HOSTS=localhost:2181 \
KM_VERSION=2.0.0.2
RUN mkdir -p /tmp && \
cd /tmp && \
wget https://github.com/yahoo/kafka-manager/archive/${KM_VERSION}.tar.gz && \
tar xf ${KM_VERSION}.tar.gz && \
cd /tmp/kafka-manager-${KM_VERSION} && \
sbt clean dist && \
unzip -d / ./target/universal/kafka-manager-${KM_VERSION}.zip && \
rm -fr /tmp/${KM_VERSION} /tmp/kafka-manager-${KM_VERSION}
WORKDIR /kafka-manager-${KM_VERSION}
EXPOSE 9000
ENTRYPOINT ["./bin/kafka-manager","-Dconfig.file=conf/application.conf"]
apiVersion: apps/v1
kind: Deployment
metadata:
name: kafka-manager
namespace: infra
labels:
name: kafka-manager
spec:
replicas: 1
selector:
matchLabels:
app: kafka-manager
template:
metadata:
labels:
app: kafka-manager
spec:
containers:
- name: kafka-manager
image: 172.16.1.10:8900/public/kafka-manager:v2.0.0.2
ports:
- containerPort: 9000
protocol: TCP
env:
- name: ZK_HOSTS
value: zk1.host.com:2181
- name: APPLICATION_SECRET
value: letmein
apiVersion: v1
kind: Service
metadata:
name: kafka-manager
namespace: infra
spec:
ports:
- protocol: TCP
port: 9000
targetPort: 9000
selector:
app: kafka-manager
apiVersion: extensions/v1beta1
kind: Ingress
metadata:
name: kafka-manager
namespace: infra
spec:
rules:
- host: kafka-manager.host.com
http:
paths:
- path: /
backend:
serviceName: kafka-manager
servicePort: 9000
安装filebeat
日志收集服务使用filebeat,因为传统的ruby收集的日志的方式十分消耗资源;其与业务服务采用边车模式运行;通过挂载卷的方式共享业务容器中的日志
ps:
- 在k8s的yaml文件中定义containers时,同时定义的容器就是以边车模式运行的
- docker的边车模式下两个容器共享了网络名称空间,USER(用户名称空间)和UTS(时间),隔离了IPC(进程空间)和文件系统
注意点:要与elasticsearch版本一致
制作镜像
# windows环境中下载好的压缩包直接ADD到镜像中 https://artifacts.elastic.co/downloads/beats/filebeat/filebeat-7.4.0-linux-x86_64.tar.gz
FROM debian:jessie
ADD filebeat-7.4.0-linux-x86_64.tar.gz /opt/
RUN set -x && cp /opt/filebeat-*/filebeat /bin && rm -fr /opt/filebeat*
COPY entrypoint.sh /
ENTRYPOINT ["/entrypoint.sh"]
filebeat]# cat docker-entrypoint.sh
#!/bin/bash
ENV=${ENV:-"dev"} # 运行环境
PROJ_NAME=${PROJ_NAME:-"no-define"} # project 名称,关系到topic
MULTILINE=${MULTILINE:-"^\d{2}"} # 多行匹配,根据日志格式来定
KAFKA_ADDR=${KAFKA_ADDR:-'"172.16.0.21:9092"'}
cat > /etc/filebeat.yaml << EOF
filebeat.inputs:
- type: log
fields_under_root: true
fields:
topic: logm-${PROJ_NAME}
paths:
- /logm/*.log
- /logm/*/*.log
- /logm/*/*/*.log
- /logm/*/*/*/*.log
- /logm/*/*/*/*/*.log
scan_frequency: 120s
max_bytes: 10485760
multiline.pattern: '$MULTILINE'
multiline.negate: true
multiline.match: after
multiline.max_lines: 100
- type: log
fields_under_root: true
fields:
topic: logu-${PROJ_NAME}
paths:
- /logu/*.log
- /logu/*/*.log
- /logu/*/*/*.log
- /logu/*/*/*/*.log
- /logu/*/*/*/*/*.log
- /logu/*/*/*/*/*/*.log
output.kafka:
hosts: [${KAFKA_ADDR}]
topic: k8s-fb-$ENV-%{[topic]}
version: 2.0.0
required_acks: 0
max_message_bytes: 10485760
EOF
set -xe
if [[ "$1" == "" ]]; then
exec filebeat -c /etc/filebeat.yaml
else
exec "$@"
fi
docker build -t 172.16.1.10:8900/filebeat:v7.4.0 .
创建docker镜像
docker push 172.16.1.10:8900/filebeat:v7.4.0
推送镜像到仓库
在spinnaker中配置日志收集容器filebeat
- 配置emptydir类型的存储卷;将/logm下的文件挂载到该配置卷下;将业务容器的日志同样挂载到该存储卷下;
- 配置环境变量,在
filebeat.yaml
配置文件中的所有环境变量都可以配置到容器环境变量中,有项目名称$PROJ_NAME
,环境名称$ENV
,多行匹配正则规则$MULTILINE
;
部署logstash
版本要求:与elasticsearch一致
启动方式:开发和生产环境各使用docker run启动一份
部署步骤:
# 下载镜像
~]# docker image pull logstash:6.8.3
~]# docker image tag logstash:6.8.3 172.16.1.10:8900/logstash:6.8.3
~]# docker image push 172.16.1.10:8900/logstash:6.8.3
cat /etc/logstash/logstash-dev.conf
input {
kafka {
bootstrap_servers => "172.16.0.21:9092"
client_id => "172.16.0.21"
consumer_threads => 4
group_id => "k8s_dev"
topics_pattern => "k8s-fb-dev-.*"
}
}
filter {
json {
source => "message"
}
}
output {
elasticsearch {
hosts => ["172.16.0.21:9200"]
index => "k8s-dev-%{+YYYY.MM}"
}
}
# 运行logstash dev环境的日志收集
docker run -d --name logstash-dev -v /etc/logstash:/etc/logstash 172.16.1.10:8900/logstash:v6.8.3 -f /etc/logstash/logstash-dev.conf
kibana部署
~]# docker pull kibana:6.8.3
~]# docker image tag kibana:6.8.3 172.16.1.10:8900/kibana:6.8.3
~]# docker image push 172.16.1.10:8900/kibana:6.8.3
pd控制器资源配置清单 dp.yml文件
apiVersion: apps/v1
kind: Deployment
metadata:
name: kibana
namespace: infra
labels:
name: kibana
spec:
replicas: 1
selector:
matchLabels:
name: kibana
template:
metadata:
labels:
app: kibana
name: kibana
spec:
containers:
- name: kibana
image: 172.16.1.10:8900/kibana:6.8.3
ports:
- containerPort: 5601
protocol: TCP
env:
- name: ELASTICSEARCH_URL
value: http://172.16.0.21:9200
service资源配置清单 svc.yml
apiVersion: v1
kind: Service
metadata:
name: kibana
namespace: infra
spec:
ports:
- protocol: TCP
port: 80
targetPort: 5601
selector:
app: kibana
ingress资源配置清单 ingress.yml
apiVersion: extensions/v1beta1
kind: Ingress
metadata:
name: kibana
namespace: infra
spec:
rules:
- host: kibana.host.com
http:
paths:
- path: /
backend:
serviceName: kibana
servicePort: 80
应用资源配置清单
添加kibana.host.com域名到bind9服务
kibana使用
-
初始化kibana,点击Explore on my own按钮;协助优化kibana,选择NO;Monitoring选项中,点击Turn On,开启监控,可以查询es的状态及索引;
-
在Management中,配置index,使用logstash中建立index的规则名称去匹配,之后选择kibana中默认的@timestamp时间戳规则,创建index patterns
-
看日志,选择Discover,选择器的使用
- 时间选择器quick,相对时间,绝对时间
- 环境选择器,对应建立的es-index,通过环境变量传递的
- 项目选择器(Filter功能的topic选项),对应topic名称,是通过环境变量传递的
- 关键字选择器;支持lusen的查询语法
-
配置常用字段
Time, message, log.file.path, hostname -
落盘的日志直接filebeat收集;没有落盘的日志,统一输出重定向
command >> /opt/logs/*.log 2>&1
更多推荐
所有评论(0)