How to Build a Centralized Rsyslog Server with ELK for Network Log Management

Introduction

As the number of servers and network devices in a data center grows, log management and querying become painful for system administrators.

Common problems faced by sysadmins

Cannot log into every server/device to view logs during routine maintenance.

Network devices have limited storage, cannot keep long‑term logs, yet issues may stem from old events.

Attackers often delete local logs to hide intrusion traces.

Monitoring tools like Zabbix cannot replace log management for events such as logins or cron jobs.

Therefore, deploying a centralized Rsyslog server is essential in the current network environment.

Advantages of Rsyslog

Most network devices support remote syslog; configuration usually only requires IP and port (default 514).

Linux servers need only a single line in /etc/rsyslog.conf to forward logs; deployment is simple.

Deployment Architecture

Rsyslog Configuration

系统环境及软件版本：CentOS Linux release 7.5.1804 (Core)
Elasticserch-6.8.4
Kibana-6.8.4
Logstash-6.8.4
Filebeat-6.8.4
Rsyslog-8.24.0

Disable SELinux

# setenforce 0
# sed -i 's#SELINUX=enforcing#SELINUX=disabled#g' /etc/selinux/config

Firewall configuration

firewall-cmd --add-service=syslog --permanent
firewall-cmd --reload

Check Rsyslog installation

CentOS 7 installs rsyslog by default.

[root@ZABBIX-Server ~]# rpm -qa |grep rsyslog
rsyslog-8.24.0-16.el7.x86_64

Edit /etc/rsyslog.conf

$ModLoad imudp
$UDPServerRun 514
$ModLoad imtcp
$InputTCPServerRun 514
$WorkDirectory /var/lib/rsyslog
$ActionFileDefaultTemplate RSYSLOG_TraditionalFileFormat
$IncludeConfig /etc/rsyslog.d/*.conf
$OmitLocalLogging on
$IMJournalStateFile imjournal.state
*.info;mail.none;authpriv.none;cron.none;local6.none;local5.none;local4.none                /var/log/messages
$template h3c,"/mnt/h3c/%FROMHOST-IP%.log"
local6.* ?h3c
$template huawei,"/mnt/huawei/%FROMHOST-IP%.log"
local5.* ?huawei
$template cisco,"/mnt/cisco/%FROMHOST-IP%.log"
local4.* ?cisco

$ModLoad imudp            # immark module, supports TCP
$ModLoad imudp            # imupd module, supports UDP
$InputTCPServerRun 514
$UDPServerRun 514          # allow UDP/TCP on port 514

Note:

*.info;mail.none;authpriv.none;cron.none;local6.none;local5.none;local4.none         /var/log/messages
By default local6.none;local5.none;local4.none are not added, so network logs are also written to /var/log/messages.

Check Rsyslog service

Restart Rsyslog

systemctl restart rsyslog.service

Log storage directories

Network devices forward logs to the syslog server; different vendors map to different local facilities:

/mnt/huawei   --- local6
/mnt/h3c      --- local5
/mnt/cisco    --- local4

Network device configuration

Huawei:
info-center loghost source Vlanif99
info-center loghost 192.168.99.50 facility local5

H3C:
info-center loghost source Vlan-interface99
info-center loghost 192.168.99.50 facility local6

CISCO:
(config)#logging on
(config)#logging 192.168.99.50
(config)#logging facility local4
(config)#logging source-interface e0

Ruijie:
logging buffered warnings
logging source interface VLAN 99
logging facility local6
logging server 192.168.99.50

Note: 192.168.99.50 is the Rsyslog server IP.

Edit Filebeat configuration

Collect logs under /mnt/* and ship them to Logstash.

filebeat.inputs:
- type: log
  enabled: true
  paths:
    - /mnt/huawei/*
  tags: ["huawei"]
  include_lines: ['Failed','failed','error','ERROR','\bDOWN\b','\bdown\b','\bUP\b','\bup\b']
  drop_fields:
    fields: ["beat","input_type","source","offset","prospector"]
- type: log
  paths:
    - /mnt/h3c/*
  tags: ["h3c"]
  include_lines: ['Failed','failed','error','ERROR','\bDOWN\b','\bdown\b','\bUP\b','\bup\b']
  drop_fields:
    fields: ["beat","input_type","source","offset","prospector"]
setup.template.settings:
  index.number_of_shards: 3
output.logstash:
  hosts: ["192.168.99.185:5044"]
processors:
  - add_host_metadata: ~
  - add_cloud_metadata: ~

Edit Logstash configuration

Parse logs from Filebeat based on tags and output to Elasticsearch.

input {
  beats {
    port => 5044
  }
}
filter {
  if "huawei" in [tags] {
    grok{
      match => {"message" => "%{SYSLOGTIMESTAMP:time} %{DATA:hostname} %{GREEDYDATA:info}"}
    }
  }
  else if "h3c" in [tags] {
    grok{
      match => {"message" => "%{SYSLOGTIMESTAMP:time} %{YEAR:year} %{DATA:hostname} %{GREEDYDATA:info}"}
    }
  }
  mutate {
    remove_field => ["message","time","year","offset","tags","path","host","@version","[log]","[prospector]","[beat]","[input][type]","[source]"]
  }
}
output{
  stdout {codec => rubydebug}
  elasticsearch {
    index => "networklogs-%{+YYYY.MM.dd}"
    hosts => ["192.168.99.185:9200"]
    sniffing => false
  }
}

Visualization in Kibana

Create an index pattern that matches the network‑device log indices.

Create a data table

Kibana tables can be exported as CSV files.

Create a pie chart

Feel free to discuss and share improvements.