如何使用logstash动态定义elasticsearch索引？

Question

参见下面的logstash配置文件，使用jdbc插件将记录从mysql数据库拉到elasticsearch索引中。如何修改它，以便根据在数据库中找到的company_id值生成单独的索引，如：company_%{company_id}_user_events。

这是否可以动态完成，或者是否需要为每个公司id预先配置和硬编码创建单独的logstash配置文件？有没有像脚本或模板这样的中间路线？

如果有帮助，可以将company_id字段添加到数据库的ahoy_events表中，而不是像现在这样通过用户关联“添加”。

当前logstash.conf

input {
    jdbc {
        jdbc_driver_library => "/opt/mysql-connector-java-5.1.47-bin.jar"
        jdbc_driver_class => "com.mysql.jdbc.Driver"
        jdbc_connection_string => "jdbc:mysql://mysql:3306/${DB_NAME}"
        jdbc_user => "${DB_USER}"
        jdbc_password => "${DB_PASSWORD}"
        schedule => "* * * * *"
        statement => "select * from ahoy_events where time > :sql_last_value"
    }
}

filter {
    jdbc_streaming {
        jdbc_driver_library => "/opt/mysql-connector-java-5.1.47-bin.jar"
        jdbc_driver_class => "com.mysql.jdbc.Driver"
        jdbc_connection_string => "jdbc:mysql://mysql:3306/${DB_NAME}"
        jdbc_user => "${DB_USER}"
        jdbc_password => "${DB_PASSWORD}"
        statement => "select * from users where id = :user"
        parameters => { "user" => "user_id" }
        target => "user"
    }
    jdbc_streaming {
        jdbc_driver_library => "/opt/mysql-connector-java-5.1.47-bin.jar"
        jdbc_driver_class => "com.mysql.jdbc.Driver"
        jdbc_connection_string => "jdbc:mysql://mysql:3306/${DB_NAME}"
        jdbc_user => "${DB_USER}"
        jdbc_password => "${DB_PASSWORD}"
        statement => "select * from visits where id = :visits"
        parameters => { "visits" => "visit_id" }
        target => "visits"
    }
    mutate {
        add_field => { "company_id" =>  "%{[user][0][company_id]}"}
    }
    jdbc_streaming {
        jdbc_driver_library => "/opt/mysql-connector-java-5.1.47-bin.jar"
        jdbc_driver_class => "com.mysql.jdbc.Driver"
        jdbc_connection_string => "jdbc:mysql://mysql:3306/${DB_NAME}"
        jdbc_user => "${DB_USER}"
        jdbc_password => "${DB_PASSWORD}"
        statement => "select * from companies where id = :company_id"
        parameters => { "company_id" => "company_id" }
        target => "company"
    }
    json {
        source => "properties"
        target => "properties"
    }
    mutate {
        add_field => { "user_name" => "%{[user][0][name]}" }
    }
    mutate {
        add_field => { "company_name" => "%{[company][0][name]}" }
    }
    mutate {
        rename => { "[visits][0]" => "visit" }
    }
    mutate {
        remove_field => ["visits", "company", "user"]
    }
}

output {
    elasticsearch {
        hosts => ["http://elasticsearch:9200"]
        index => "user_events-%{+YYYY.MM.dd}"
        document_id => "%{id}"
    }
}

所需的结果是一个名称空间为company_id：company_%{company_id}_user_events的索引，这样我以后就可以在相同模式的company_%{company_id}_other_records中添加其他索引

Answer 1

不是100%确定，但从技术上讲应该就这么简单：

output {
    elasticsearch {
        hosts => ["http://elasticsearch:9200"]
        index => "company_%{company_id}_events-%{+YYYY.MM.dd}"
        document_id => "%{id}"
    }
}