目录
一、背景
二、设计
三、具体实现
Filebeat配置
K8S SideCar yaml
Logstash配置
一、背景
将容器中服务的trace日志和应用日志收集到KAFKA，需要注意的是 trace 日志和app 日志需要存放在同一个KAFKA两个不同的topic中。分别为APP_TOPIC和TRACE_TOPIC
二、设计
流程图如下：
说明：
APP_TOPIC：主要存放服务的应用日志
TRACE_TOPIC：存放程序输出的trace日志，用于排查某一个请求的链路
文字说明：
filebeat 采集容器中的日志（这里需要定义一些规范，我们定义的容器日志路径如下）,filebeat会采集两个不同目录下的日志，然后输出到对应的topic中，之后对kafka 的topic进行消费、存储。最终展示出来
/home/service/
└── logs
├── app
│ └── pass
│ ├── 10.246.84.58-paas-biz-784c68f79f-cxczf.log
│ ├── 1.log
│ ├── 2.log
│ ├── 3.log
│ ├── 4.log
│ └── 5.log
└── trace
├── 1.log
├── 2.log
├── 3.log
├── 4.log
├── 5.log
└── trace.log
4 directories, 13 files三、具体实现
上干货～
Filebeat配置
配置说明：
其中我将filebeat的一些配置设置成了变量，在接下来的k8s yaml文件中需要定义变量和设置变量的value。
需要特别说明的是我这里是使用了 tags: ["trace-log"]结合when.contains来匹配，实现将对应intput中的日志输出到对应kafka的topic中
filebeat.inputs:
- type: log
enabled: true
paths:
- /home/service/logs/trace/*.log
fields_under_root: true
fields:
topic: "${TRACE_TOPIC}"
json.keys_under_root: true
json.add_error_key: true
json.message_key: message
scan_frequency: 10s
max_bytes: 10485760
harvester_buffer_size: 1638400
ignore_older: 24h
close_inactive: 1h
tags: ["trace-log"]
processors:
- decode_json_fields:
fields: ["message"]
process_array: false
max_depth: 1
target: ""
overwrite_keys: true
- type: log
enabled: true
paths:
- /home/service/logs/app/*/*.log
fields:
topic: "${APP_TOPIC}"
scan_frequency: 10s
max_bytes: 10485760
harvester_buffer_size: 1638400
close_inactive: 1h
tags: ["app-log"]
output.kafka:
enabled: true
codec.json:
pretty: true # 是否格式化json数据，默认false
compression: gzip
hosts: "${KAFKA_HOST}"
topics:
- topic: "${TRACE_TOPIC}"
bulk_max_duration: 2s
bulk_max_size: 2048
required_acks: 1
max_message_bytes: 10485760
when.contains:
tags: "trace-log"
- topic: "${APP_TOPIC}"
bulk_flush_frequency: 0
bulk_max_size: 2048
compression: gzip
compression_level: 4
group_id: "k8s_filebeat"
grouping_enabled: true
max_message_bytes: 10485760
partition.round_robin:
reachable_only: true
required_acks: 1
workers: 2
when.contains:
tags: "app-log"K8S SideCar yaml
配置说明：
该yaml中定一个两个容器，容器1为nginx（示例）容器2为filebeat容器。定义了一个名称为logs的emptryDir类型的卷，将logs卷同时挂载在了容器1和容器2的/home/service/logs目录
接下来又在filebeat容器中自定义了三个环境变量，这样我们就可以通过修改yaml的方式很灵活的来配置filebeat
TRACE_TOPIC： Trace日志的topic
APP_TOPIC：App日志的topic
KAFKA_HOST：KAFKA地址
apiVersion: apps/v1
kind: Deployment
metadata:
labels:
app: nginx
name: nginx
namespace: default
spec:
replicas: 2
selector:
matchLabels:
app: nginx
template:
metadata:
labels:
app: nginx
spec:
imagePullSecrets:
- name: uhub-registry
containers:
- image: uhub.service.ucloud.cn/sre-paas/nginx:v1
imagePullPolicy: IfNotPresent
name: nginx
ports:
- name: nginx
containerPort: 80
- mountPath: /home/service/logs
name: logs
terminationMessagePath: /dev/termination-log
terminationMessagePolicy: File
volumeMounts:
- mountPath: /home/service/logs
name: logs
- env:
- name: TRACE_TOPIC
value: pro_platform_monitor_log
- name: APP_TOPIC
value: platform_logs
- name: KAFKA_HOST
value: '["xxx.xxx.xxx.xxx:9092","xx.xxx.xxx.xxx:9092","xx.xxx.xxx.xxx:9092"]'
- name: MY_POD_NAME
valueFrom:
fieldRef:
apiVersion: v1
fieldPath: metadata.name
image: xxx.xxx.xxx.cn/sre-paas/filebeat-v2:8.11.2
imagePullPolicy: Always
name: filebeat
resources:
limits:
cpu: 150m
memory: 200Mi
requests:
cpu: 50m
memory: 100Mi
securityContext:
privileged: true
runAsUser: 0
terminationMessagePath: /dev/termination-log
terminationMessagePolicy: File
volumeMounts:
- mountPath: /home/service/logs
name: logs
dnsPolicy: ClusterFirst
imagePullSecrets:
- name: xxx-registry
restartPolicy: Always
schedulerName: default-scheduler
securityContext: {}
terminationGracePeriodSeconds: 30
volumes:
- emptyDir: {}
name: logs Logstash配置
input {
kafka {
type => "platform_logs"
bootstrap_servers => "xxx.xxx.xxx.xxx:9092,xxx.xxx.xxx.xxx:9092,xxx.xxx.xxx.xxx:9092"
topics => ["platform_logs"]
group_id => 'platform_logs'
client_id => 'open-platform-logstash-logs'
}
kafka {
type => "platform_pre_log"
bootstrap_servers => "xxx.xxx.xxx.xxx:9092,xxx.xxx.xxx.xxx:9092,xxx.xxx.xxx.xxx:9092"
topics => ["pre_platform_logs"]
group_id => 'pre_platform_logs'
client_id => 'open-platform-logstash-pre'
}
kafka {
type => "platform_nginx_log"
bootstrap_servers => "xxx.xxx.xxx.xxx:9092,xxx.xxx.xxx.xxx:9092,xxx.xxx.xxx.xxx:9092"
topics => ["platform_nginx_log"]
group_id => 'platform_nginx_log'
client_id => 'open-platform-logstash-nginx'
}
}
filter {
if [type] == "platform_pre_log" {
grok {
match => { "message" => "\[%{IP}-(?<service>[a-zA-Z-]+)-%{DATA}\]" }
}
}
if [type] == "platform_logs" {
grok {
match => { "message" => "\[%{IP}-(?<service>[a-zA-Z-]+)-%{DATA}\]" }
}
}
}
output {
if [type] == "platform_logs" {
elasticsearch {
id => "platform_logs"
hosts => ["http://xxx.xxx.xxx.xxx:9200","http://xxx.xxx.xxx.xxx:9200","http://xxx.xxx.xxx.xxx:9200"]
index => "log-xxx-prod-%{service}-%{+yyyy.MM.dd}"
user => "logstash_transformer"
password => "xxxxxxx"
template_name => "log-xxx-prod"
manage_template => "true"
template_overwrite => "true"
}
}
if [type] == "platform_pre_log" {
elasticsearch {
id => "platform_pre_logs"
hosts => ["http://xxx.xxx.xxx.xxx:9200","http://xxx.xxx.xxx.xxx:9200","http://xxx.xxx.xxx.xxx:9200"]
index => "log-xxx-pre-%{service}-%{+yyyy.MM.dd}"
user => "logstash_transformer"
password => "xxxxxxx"
template_name => "log-xxx-pre"
manage_template => "true"
template_overwrite => "true"
}
}
if [type] == "platform_nginx_log" {
elasticsearch {
id => "platform_nginx_log"
hosts => ["http://xxx.xxx.xxx.xxx:9200","http://xxx.xxx.xxx.xxx:9200","http://xxx.xxx.xxx.xxx:9200"]
index => "log-platform-nginx-%{+yyyy.MM.dd}"
user => "logstash_transformer"
password => "xxxxxxx"
template_name => "log-platform-nginx"
manage_template => "true"
template_overwrite => "true"
}
}
}如果有帮助到你麻烦给个赞或者收藏一下～，有问题可以随时私聊我或者在评论区评论，我看到会第一时间回复
![[文本挖掘和知识发现] 03.基于大连理工情感词典的情感分析和情绪计算](https://img-blog.csdnimg.cn/20200809125022536.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L0Vhc3Rtb3VudA==,size_16,color_FFFFFF,t_70#pic_center)
