This parser is deprecated. Use Looking up GeoIP2 data from IP addresses instead.
The syslog-ng OSE application can lookup IPv4 addresses from an offline GeoIP database, and make the retrieved data available in name-value pairs. IPv6 addresses are not supported. Depending on the database used, you can access country code, longitude, and latitude information.
NOTE: To access longitude and latitude information, download the GeoLiteCity database, and unzip it (for example, to the /usr/share/GeoIP/GeoLiteCity.dat file). The default databases available on Linux and other platforms usually contain only the country codes.
You can refer to the separated parts of the message using the key of the value as a macro. For example, if the message contains KEY1=value1,KEY2=value2, you can refer to the values as ${KEY1} and ${KEY2}.
Declaration:
parser parser_name {
geoip(
<macro-containing-the-IP-address-to-lookup>
prefix()
database("<path-to-database-file>")
);
};Example: Using the GeoIP parser
In the following example, syslog-ng OSE retrieves the GeoIP data of the IP address contained in the ${HOST} field of the incoming message, and includes the data (prefixed with the geoip. string) in the output JSON message.
@version: 3.7
options {
keep-hostname(yes);
};
source s_file {
file("/tmp/input");
};
parser p_geoip { geoip( "${HOST}", prefix( "geoip." ) database( "/usr/share/GeoIP/GeoLiteCity.dat" ) ); };
destination d_file {
file( "/tmp/output" template("$(format-json --scope core --key geoip*)\n") );
};
log {
source(s_file);
parser(p_geoip);
destination(d_file);
};
For example, for the <38>Jan 1 14:45:22 192.168.1.1 prg00000[1234]: test message message the output will look like:
{"geoip":{"longitude":"47.460704","latitude":"19.049968","country_code":"HU"},"PROGRAM":"prg00000","PRIORITY":"info","PID":"1234","MESSAGE":"test message","HOST":"192.168.1.1","FACILITY":"auth","DATE":"Jan 1 14:45:22"}
If you are transferring your log messages into Elasticsearch, use the following rewrite rule to combine the longitude and latitude information into a single value (called geoip.location), and set the mapping in Elasticsearch accordingly. Do not forget to include the rewrite in your log path. For details on transferring your log messages to Elasticsearch, see elasticsearch2: Sending messages directly to Elasticsearch version 2.0 or higher (DEPRECATED).
rewrite r_geoip {
set(
"${geoip.latitude},${geoip.longitude}",
value( "geoip.location" ),
condition(not "${geoip.latitude}" == "")
);
};
In your Elasticsearch configuration, set the appropriate mappings:
{
"mappings" : {
"_default_" : {
"properties" : {
"geoip" : {
"properties" : {
"country_code" : {
"index" : "not_analyzed",
"type" : "string",
"doc_values" : true
},
"latitude" : {
"index" : "not_analyzed",
"type" : "string",
"doc_values" : true
},
"longitude" : {
"type" : "string",
"doc_values" : true,
"index" : "not_analyzed"
},
"location" : {
"type" : "geo_point"
}
}
}
}
}
}
}