NOTE: This project is no longer being maintained.

Fluentd filter plugin to transform records.

This gem provides the mutate filter for Fluentd which is designed to replicate the way mutate works in Logstash.

$ gem install fluent-plugin-mutate_filter

Add following line to your Gemfile:

gem "fluent-plugin-mutate"

And then execute:

$ bundle

The mutate filter accepts a list of <mutate> blocks which will be executed in sequence, applying their respective mutation @type.

Each block will be evaluated in the order with which they were defined in the configuration file. Multiple keys within a same block should be evaluated in the defined order, however this shouldn't be relied upon.

While evaluating mutations, should an exception occur, it will be logged and further mutations will continue to be evaluated.

<filter **>
  @type mutate

  <mutate>
    @type     rename
    old_key1  "new_key1"
    old_key2  "new_key2"
  </mutate>

  <mutate>
    @type     gsub
    new_key2  ["^\\[log \\d{4}-\\d{2}-\\d{2}T.*\\] {","{"]
  </mutate>
</filter

Global options:

• expand_nesting (true)|false: Treat periods in field names as separators. Allowing you to reference { deeply: { nested: { key: value } } } as deeply.nested.key.
• prune_empty (true)|false: At the end of the list of mutations, delete any values with empty Maps or Arrays or Nil values.

Certain mutators, such as gsub, replace and update may contain special patterns which will be evaluated during the mutation. Supported patterns are:

• %{key}: Pattern will be replaced by record key value. When expand_nesting is enabled, the key may be in dot notation.
• %e{ENV_VAR}: Pattern will be replaced by an environment variable. If ENV\_VAR is hostname, then the machine's hostname will be injected.

Convert the key's value to a given type.

Apply a regex find and replace on a key's value.

Additionally, new_value may contain replacement patterns which will be evaluated during the replacement.

Join the components of an Array key using separator, converting it to a String.

Lowercase the key's value. If the value is an Array, then apply the filter to all array elements.

Merge the value of source_field, which may be an Array or a Map, into the key's value.

Parse a key's value from JSON to a object.

Additionally, a @merge_root boolean field may be defined to have the resulting object be merge into the root of the record.

Rename a key to a new key name.

Remove a key and it's value, whether the value be a string or an object.

Replace the value of a key with new_value, if the key exists. Otherwise, set the value of the key to new_value.

Additionally, new_value may contain replacement patterns which will be evaluated during the replacement.

Replace the value of a key with new_value, if the key exists. Otherwise, ignore this operation.

Additionally, new_value may contain replacement patterns which will be evaluated during the replacement.

Uppercase the key's value. If the value is an Array, then apply the filter to all array elements.

Split the value of key using separator, converting it to an Array.

Remove whitespace surrounding a key's value. If the value is an Array, apply the filter to each element.

• Copyright(c) 2018- Jonathan Serafini
• License
  • Apache License, Version 2.0