VitePress

Appearance

Return to top
On this page

Kafka + ELFK分布式日志收集

Filebeat --> Kafka --> Logstash --> ES Cluster --> Kibana/Grafana

使用背景

由于ELFK的局限性，随着 Beats 收集的每秒数据量越来越大，Logstash 可能无法承载这么大量日志的处理。虽然说，可以增加 Logstash 节点数量，提高每秒数据的处理速度，但是仍需考虑可能Elasticsearch无法承载这么大量的日志的写入。此时，我们可以考虑 引入消息队列 （Kafka），进行缓存。

Filebeat --> Kafka 

# vim /usr/local/filebeat-7.7.1-linux-x86_64/filebeat.yml
filebeat.inputs:
- type: log
  paths:
    - /usr/local/apache-tomcat-9.0.34/logs/tomcat_access_log.*.log

output.kafka:
  # initial brokers for reading cluster metadata
  #hosts: ["kafka1:9092", "kafka2:9092", "kafka3:9092"]
  hosts: ["172.16.1.131:9092"]
  # message topic selection + partitioning
  # topic: '%{[fields.log_topic]}'
  topic: tomcat-log
  partition.round_robin:
    reachable_only: false

# vim /usr/local/filebeat-7.7.1-linux-x86_64/filebeat.yml
filebeat.inputs:
- type: log
  paths:
    - /usr/local/apache-tomcat-9.0.34/logs/tomcat_access_log.*.log

output.kafka:
  # initial brokers for reading cluster metadata
  #hosts: ["kafka1:9092", "kafka2:9092", "kafka3:9092"]
  hosts: ["172.16.1.131:9092"]
  # message topic selection + partitioning
  # topic: '%{[fields.log_topic]}'
  topic: tomcat-log
  partition.round_robin:
    reachable_only: false

启动 Kafka

这里为了方便起见，我使用docker-compose脚本来进行快速部署，脚本内容参见[使用 docker-compose 部署 Kafka](../Kafka/使用 docker-compose 部署 Kafka.md)

Logstash --> ElasticSearch 

# vim /usr/local/logstash-7.7.1/config/kafka2es.conf
input {
    kafka {
    codec => "json"
    topics => ["tomcat-log"]
    bootstrap_servers => "172.16.1.131:9092"
    auto_offset_reset => "latest"
    group_id => "logstash-g1"
    }
}
output {
    elasticsearch {
    hosts => "http://172.16.1.131:9200"
    index => "filebeat_%{[fields][log_source]}-%{+YYYY.MM.dd}"
	}
}

# vim /usr/local/logstash-7.7.1/config/kafka2es.conf
input {
    kafka {
    codec => "json"
    topics => ["tomcat-log"]
    bootstrap_servers => "172.16.1.131:9092"
    auto_offset_reset => "latest"
    group_id => "logstash-g1"
    }
}
output {
    elasticsearch {
    hosts => "http://172.16.1.131:9200"
    index => "filebeat_%{[fields][log_source]}-%{+YYYY.MM.dd}"
	}
}