I've put together a moderately-sized rewrite based heavily on this repository that adds support for multiple MMDB formats, as well as the ability to iterate over a list of source_addr_fields in a Heka message. Any interest in a PR submission? Thanks!

https://github.com/robison/heka-geoip2-decoder

Hi there,

i used your instructions and compiled heka with geoip2-decoder plugin. now i am trying to decode the user location based on ip's from nginx access log but I have no success.

Here is my toml file

[Nginx-Webserver-Access]
type = "LogstreamerInput"
log_directory = "/home/guest"
file_match = 'test.log'
decoder = "shipped-nginx-decoder"

[shipped-nginx-decoder]
type = "MultiDecoder"
subs = ['CombinedLogDecoder', 'GeoIp2Decoder']

[CombinedLogDecoder]
type = "SandboxDecoder"
filename = "lua_decoders/nginx_access.lua"

[CombinedLogDecoder.config]
type = "nginx.access"
user_agent_transform = true

combined log format

log_format = '$remote_addr - $remote_user [$time_local] "$request" $status $body_bytes_sent "$http_referer" "$http_user_agent"'
payload_keep = true

[GeoIp2Decoder]
db_file = "/home/guest/GeoLite2-City.mmdb"
source_host_field = "remote_addr"
target_field_prefix = "geoip"
dns_lookup = false

[ESJsonEncoder]
index = "test-%{2006.01.02}"
es_index_from_timestamp = true
type_name = "%{Type}"
fields = ["Timestamp","Payload","HostName","Fields"]

[LogOutput]
message_matcher = "Type == 'nginx.access'"
encoder = "ESJsonEncoder"

Please let me know if am doing right way.