Reputation: 3172
i'm very new to logstash and elastic search. I am trying to store log files both in elasticsearch and a flat file. I know that logstash support both output. But are they processed simultaneously? or is it done periodically through a job?
Upvotes: 6
Views: 24104
Reputation: 388
First you need to install output plugins:
/usr/share/logstash/bin/logstash-plugin install logstash-output-elasticsearch
/usr/share/logstash/bin/logstash-plugin install logstash-output-file
Then create conf files for output:
cat /etc/logstash/conf.d/nfs-output.conf
output {
file {
path => "/your/path/filebeat-%{+YYYY-MM-dd}.log"
}
}
cat /etc/logstash/conf.d/30-elasticsearch-output.conf
output {
elasticsearch {
hosts => ["elasitc_ip:9200"]
manage_template => true
user => "elastic"
password => "your_password"
}
}
Then:
service logstash restart
Upvotes: 0
Reputation: 1043
Yes you can do this like so by tagging and cloning your inputs with the "add_tag" command on your shipper config.
input
{
tcp { type => "linux" port => "50000" codec => plain { charset => "US-ASCII" } }
tcp { type => "apache_access" port => "50001" codec => plain { charset => "US-ASCII" } }
tcp { type => "apache_error" port => "50002" codec => plain { charset => "US-ASCII" } }
tcp { type => "windows_security" port => "50003" codec => plain { charset => "US-ASCII" } }
tcp { type => "windows_application" port => "50004" codec => plain { charset => "US-ASCII" } }
tcp { type => "windows_system" port => "50005" codec => plain { charset => "US-ASCII" } }
udp { type => "network_equipment" port => "514" codec => plain { charset => "US-ASCII" } }
udp { type => "firewalls" port => "50006" codec => plain }
}
filter
{
grok { match => [ "host", "%{IPORHOST:ipaddr}(:%{NUMBER})?" ] }
mutate { replace => [ "fqdn", "%{ipaddr}" ] }
dns { reverse => [ "fqdn", "fqdn" ] action => "replace" }
if [type] == "linux" { clone { clones => "linux.log" add_tag => "savetofile" } }
if [type] == "apache_access" { clone { clones => "apache_access.log" add_tag => "savetofile" } }
if [type] == "apache_error" { clone { clones => "apache_error.log" add_tag => "savetofile" } }
if [type] == "windows_security" { clone { clones => "windows_security.log" add_tag => "savetofile" } }
if [type] == "windows_application" { clone { clones => "windows_application.log" add_tag => "savetofile" } }
if [type] == "windows_system" { clone { clones => "windows_system.log" add_tag => "savetofile" } }
if [type] == "network_equipment" { clone { clones => "network_%{fqdn}.log" add_tag => "savetofile" } }
if [type] == "firewalls" { clone { clones => "firewalls.log" add_tag => "savetofile" } }
}
output
{
#stdout { debug => true }
#stdout { codec => rubydebug }
redis { host => "1.1.1.1" data_type => "list" key => "logstash" }
}
And on your main logstash instance you would do this:
input {
redis {
host => "1.1.1.1"
data_type => "list"
key => "logstash"
type=> "redis-input"
# We use the 'json' codec here because we expect to read json events from redis.
codec => json
}
}
output
{
if "savetofile" in [tags] {
file {
path => [ "/logs/%{fqdn}/%{type}" ] message_format => "%{message}"
}
}
else { elasticsearch { host => "2.2.2.2" }
}
}
Upvotes: 6
Reputation: 7890
FYI, You can study The life of logstash event about the logstash event.
The output worker model is currently a single thread. Outputs will receive events in the order they are defined in the config file.
But the Outputs may decide to buffer events temporarily before publishing them. Ex: Output will buffers 2 or 3 events then just it write to file.
Upvotes: 0