I was wondering how to use env vars in the Fluentd config, I tried:
<match **>
type elasticsearch
logstash_format true
logstash_prefix $ENV_VAR
host ***
port ***
include_tag_key true
tag_key _key
</match>
but it doesn't work, any idea?
EDIT:
Here is a much better solution:
If you pass "--use-v1-config" option to Fluentd, this is possible with the "#{ENV['env_var_name']" like this:
<match foobar.**> # ENV["FOO"] is foobar
type elasticsearch
logstash_prefix "#{ENV['FOO']}"
logstash_format true
include_tag_key true
tag_key _key
host ****
port ****
</match>
Old, kludgey answer is here.
fluent-plugin-record-reformer
and fluent-plugin-forest
<match hello.world>
type record_reformer
tag ${ENV["FOO"]}.${tag_prefix[-1]} # adding the env variable as a tag prefix
</match>
<match foobar.**> # ENV["FOO"] is foobar
type forest
subtype elasticsearch
<template>
logstash_prefix ${tag_parts[0]}
logstash_format true
include_tag_key true
tag_key _key
host ****
port ****
</template>
</match>
In particular, do NOT use <match **>
there. That would catch all events and will lead to behaviors that are hard to debug.